fix M

.gitignore

@@ -1,7 +1,6 @@
 __pycache__/
 *.py[cod]
 junk/
-.vscode
 .ipynb_checkpoints/
 tools/ansible/inventory/hawaii.yml
 tools/ansible/inventory/peppermint.yml
@@ -17,6 +16,7 @@ dist/
 *.egg-info/
 .venv/
 venv/
+docs/_build/
 .codex
 
 
@@ -24,4 +24,4 @@ venv/
 *.webm
 *.mp4
 ffmpeg2pass-0.log
-*.sup
+*.sup

.vscode/extensions.json

@@ -0,0 +1,11 @@
+{
+  "recommendations": [
+    "swyddfa.esbonio",
+    "ms-python.python",
+    "ms-python.vscode-pylance",
+    "ms-python.debugpy",
+    "tamasfe.even-better-toml",
+    "redhat.vscode-yaml",
+    "DavidAnson.vscode-markdownlint"
+  ]
+}

.vscode/settings.json

@@ -0,0 +1,18 @@
+{
+  "esbonio.sphinx.pythonCommand": "${venv:.venv}/bin/python",
+  "esbonio.sphinx.buildCommand": [
+    "sphinx-build",
+    "-b",
+    "html",
+    "docs",
+    "docs/_build/html"
+  ],
+  "python.defaultInterpreterPath": "${workspaceFolder}/.venv/bin/python",
+  "python.testing.pytestEnabled": true,
+  "python.testing.pytestArgs": [
+    "--ignore=tests/legacy",
+    "--ignore=tests/support",
+    "tests"
+  ],
+  "restructuredtext.confPath": "${workspaceFolder}/docs"
+}

AGENTS.md

@@ -0,0 +1,376 @@
+# AGENTS.md
+
+This file is the entry point for agent guidance in this repository.
+
+It is intentionally generic and reusable across projects. Keep this file focused on non-project-specific constraints, working style, and the structure used to link more detailed guidance.
+
+# Purpose
+
+- Provide a small default rule set for agents working in this repository.
+- Keep the base guidance modular and easy to extend.
+- Separate reusable agent behavior from project-specific requirements.
+
+# Comment Syntax
+
+- A segment wrapped in `<!--` and `-->` is a comment and must be ignored by agents.
+- Use HTML comments for optional guidance that should stay inactive until enabled.
+- To enable an optional segment, remove the surrounding `<!--` and `-->` markers.
+
+# Core Principles
+
+- Prefer the simplest solution that satisfies the current goal.
+- Keep guidance lightweight: only add detail when it meaningfully improves outcomes.
+- Reuse modular guideline files instead of expanding this file indefinitely.
+- Treat project-specific documents as the source of truth for project behavior.
+- When guidance conflicts, use the most specific applicable document.
+
+# Rule Terms
+
+- A `rule` is the general term for any constraint, requirement, definition, or similar guidance item.
+- A `rule set` addresses all rules inside one file that share the same rule set ID.
+- Any rule inside a rule set shall use an ID following the schema `RULESET-0001`, `RULESET-0002`, and so on.
+- Rules without a rule set ID are also valid, but they are not addressable by rule ID.
+
+# Scope Of This File
+
+This file should contain:
+
+- Generic agent behavior and constraints.
+- Rules that are reusable across multiple projects.
+- Links to optional guideline modules.
+- Links to project-specific requirements.
+- Commented optional templates for released-product documentation and agent-output locations.
+
+This file should not contain:
+
+- Project business requirements.
+- Project architecture decisions.
+- Stack-specific implementation details unless they are universally applicable.
+- Task-specific runbooks that belong in dedicated modules.
+
+# Default Agent Behavior
+
+- Read the relevant context before making changes.
+- Prefer small, understandable edits over broad refactors.
+- Preserve existing patterns unless there is a clear reason to change them.
+- Document assumptions when context is missing.
+- Ignore HTML comment segments.
+- If a more specific enabled guideline exists for the current task, follow it.
+
+# Guideline Structure
+
+Use the following structure for reusable guidance files and project-specific documentation as needed:
+
+```text
+/
+|-- AGENTS.md
+|-- guidance/
+|   |-- stacks/
+|   |-- conventions/
+|   `-- workflows/
+|-- prompts/
+`-- requirements/
+
+Optional files and directories
+|-- SCRATCHPAD.md
+|-- docs/
+|   |-- readme.md
+|   |-- installation.md
+|   `-- history.md
+|-- process/
+|   |-- log.md
+|   `-- coding-handbook.md
+```
+
+# Optional Reusable Modules
+
+Add files under `guidance/` only when they are needed.
+
+# Optional Scratchpad
+
+- `SCRATCHPAD.md` is an optional repo-root scratchpad for temporary
+  information aimed at the next iteration.
+- Developers may create or delete `SCRATCHPAD.md` at any time.
+- Developers may refer to `SCRATCHPAD.md` as `scratchpad` when giving agents a
+  source or target for information.
+- Agents may read, update, create, or remove the scratchpad when the task
+  explicitly calls for it.
+- Treat the scratchpad as low-formality working context rather than canonical
+  project truth.
+- Use the scratchpad for short-lived notes, open questions, sketches, and
+  temporary decisions that should be resolved away.
+- Move durable outcomes into `requirements/`, `guidance/`, code, tests, or
+  another long-lived location.
+- If `SCRATCHPAD.md` is absent, agents should continue normally.
+
+# Optional Rule Sets
+
+- Optional rule sets may be stored in `guidance/optional/` or in `guidance/{section}/optional/`.
+- Optional rule sets are inactive by default and shall only be applied when a prompt explicitly requests them, for example by phrases such as `Apply rules for lean interface iteration in the following steps.` or `Apply LII rules.`
+- An optional rule set may be requested by its descriptive name, by its rule set ID, or by another equally clear explicit reference.
+- Agents shall never infer or auto-enable optional rule sets from general intent alone.
+- If an optional rule or rule set cannot be identified and addressed clearly, agents shall stop and ask before proceeding.
+
+# Prepared Orders
+
+- An `order` is a prepared prompt for one isolated operation rather than a general workflow or standing rule set.
+- Orders shall be stored under `prompts/`.
+- Order files shall use the naming schema `ORDER-0001-<slug>.md`, `ORDER-0002-<slug>.md`, and so on.
+- The canonical order identifier is the `ORDER-0001` style prefix. The trailing slug is descriptive only.
+- Recommended internal order file structure is: prompt ID, prompt name, purpose, trigger examples, scope, operation, and expected output.
+- Orders shall only be executed when they are explicitly requested by a prompt such as `Execute ORDER-0007.` or `Execute ORDER 7.`
+- Agents may accept an unambiguous short numeric reference such as `ORDER 7` as an alias for `ORDER-0007`.
+- If an order cannot be identified uniquely and clearly, agents shall stop and ask before proceeding.
+
+# Toolstack Guides
+
+Location:
+
+```text
+guidance/stacks/
+```
+
+Examples:
+
+- `guidance/stacks/python.md`
+- `guidance/stacks/typescript.md`
+- `guidance/stacks/docker.md`
+- `guidance/stacks/terraform.md`
+
+Use for:
+
+- Language or framework expectations.
+- Tooling and environment conventions.
+- Build, test, and runtime guidance tied to a specific stack.
+
+# Coding Conventions
+
+Location:
+
+```text
+guidance/conventions/
+```
+
+Examples:
+
+- `guidance/conventions/naming.md`
+- `guidance/conventions/testing.md`
+- `guidance/conventions/review.md`
+
+Use for:
+
+- Naming and structure conventions.
+- Testing expectations.
+- Code review and quality rules.
+
+# Recurring Workflows
+
+Location:
+
+```text
+guidance/workflows/
+```
+
+Examples:
+
+- `guidance/workflows/feature-delivery.md`
+- `guidance/workflows/bugfix.md`
+- `guidance/workflows/release.md`
+- `guidance/workflows/incident-response.md`
+
+Use for:
+
+- Repeatable task flows.
+- Checklists for common delivery work.
+- Operational or maintenance procedures.
+
+
+<!-- Enable this optional section by removing the outer HTML comment markers from this segment
+when you want agents to create, update, and consult released-product
+documentation in `docs/`.
+
+# Released Product Documentation
+ 
+Released-product documentation should live outside the generic sections above.
+
+Recommended location:
+
+```text
+docs/
+```
+
+Examples:
+
+- `docs/readme.md`
+- `docs/installation.md`
+- `docs/history.md`
+
+Agent rules for docs output:
+
+- Keep content compact but comprehensive.
+- Write for end users, operators, or other consumers of the released product.
+- Prefer shipped behavior, supported workflows, and stable terminology over
+  internal implementation detail.
+- Keep documentation synchronized with released behavior.
+- Update release history when user-visible changes are shipped.
+
+Recommended topics:
+
+- Product overview and intended use.
+- Installation, configuration, and upgrade guidance.
+- Usage patterns, operational instructions, and support boundaries.
+- Compatibility notes, migration notes, and release history.
+- Troubleshooting and common pitfalls when relevant. -->
+
+
+<!-- Enable this optional section by removing the outer HTML comment markers from this
+segment when you want agents to produce and consult workflow output in `process/`.
+
+# Agent Output In `process/`
+
+The `process/` directory is primarily for agent output created during
+delivery, maintenance, and review work.
+
+Recommended location:
+
+```text
+process/
+```
+
+Agent rules for process output:
+
+- Use `process/` for agent-produced artifacts rather than released-product
+  documentation.
+- Keep entries concise, traceable, and tied to resulting changes.
+- Treat `process/` as workflow output, not as the primary source of product
+  truth.
+- Prefer summaries and rationale over raw transcript dumps unless a workflow
+  explicitly requires full prompt history.
+
+# Agent Change Log
+
+Location:
+
+```text
+process/log.md
+```
+
+Use for:
+
+- Capturing prompts given to agents.
+- Recording concise explanations of the resulting changes made by agents.
+- Preserving task-by-task rationale, decisions, and implementation notes.
+
+# Coding Handbook
+
+Location:
+
+```text
+process/coding-handbook.md
+```
+
+Use for:
+
+- A tutorial-style handbook that explains the programming components used in
+  the project.
+- Compact but comprehensive technical onboarding material for future
+  contributors.
+- Written explanations that connect code structure, concepts, and
+  implementation patterns. -->
+
+
+# Project-Specific Requirements
+
+
+Project-specific material should live outside the generic sections above.
+
+Recommended location:
+
+```text
+requirements/
+```
+
+Examples:
+
+- `requirements/project.md`
+- `requirements/architecture.md`
+- `requirements/decisions.md`
+- `requirements/domain.md`
+
+Use for:
+
+- Product and business requirements.
+- Project goals and constraints.
+- Architecture and design decisions.
+- Domain knowledge that is specific to this repository.
+
+# Agent-Level Variables
+
+When present, `requirements/identifiers.yml` is an optional project-specific
+input that defines agent-level variables for use inside `requirements/` and
+`guidance/`.
+
+Variable schema:
+
+- Use `@{VARIABLE_NAME}` for agent-level variables.
+- Prefer uppercase snake case names such as `@{PROJECT_ID}` or `@{VENDOR_ID}`.
+- Do not treat `${...}` as an agent-level variable form; that syntax may appear
+  in Bash or other code and should not be interpreted as agent metadata.
+
+Scope:
+
+- The effective scope of `requirements/identifiers.yml` is limited to
+  `requirements/` and `guidance/`.
+- Definitions from `requirements/identifiers.yml` must not leak into product code.
+
+Defaults:
+
+- Default `@{VENDOR_ID}` is `osgw`.
+- Default `@{PROJECT_ID}` is the current repository directory name.
+
+Resolution rules:
+
+- Treat `requirements/identifiers.yml` as optional; when it is absent, agents
+  may still resolve the defaults defined above.
+- If a variable is used in `requirements/` or `guidance/` and it is not
+  defined in `requirements/identifiers.yml` and does not have a default in this
+  file, agents may stop and report the undefined variable.
+- Prefer updating duplicated identifier values in `requirements/` and
+  `guidance/` to use the variable schema when that improves consistency.
+
+# Precedence
+
+Some precedence levels may be absent because optional levels can remain inside
+HTML comments. The smaller numeric index wins.
+
+Apply guidance in this order:
+
+1. Direct user or task instructions.
+2. Project-specific documents in `requirements/`.
+<!-- 3. Released-product documentation in `docs/` when shipped behavior or
+user-facing expectations are relevant. -->
+4. Relevant modular guides in `guidance/stacks/`, `guidance/conventions/`, or `guidance/workflows/`.
+<!-- 5. Agent output in `process/` when prior prompts, rationale, or
+implementation notes are relevant. -->
+6. This `AGENTS.md`.
+
+# Maintenance
+
+- Keep this file short and stable.
+- Move detail into dedicated modules when a section becomes too specific or too long.
+- Add new guideline files only when they solve a recurring need.
+- Remove outdated references when the repository structure changes.
+
+# Current Status
+
+This repository defines the base `AGENTS.md` structure plus project-specific
+requirements and modular guidance.
+
+Future project work can add:
+
+- Reusable modules under `guidance/`
+- Project-specific documentation under `requirements/`
+- Optional temporary iteration context in `SCRATCHPAD.md`
+- Optional released-product documentation under `docs/` by uncommenting its segment
+- Optional agent output under `process/` by uncommenting its segment
+- Cross-references from this file once those documents exist

SCRATCHPAD.md

@@ -0,0 +1,75 @@
+# Scratchpad
+
+## Goal
+
+- Capture a compact, project-wide list of optimization candidates after a broad scan of the current FFX codebase, tooling, and requirements.
+
+## Focused Snapshot
+
+- Highest-leverage application optimizations:
+  - Decide whether placeholder help/settings screens should ship or disappear.
+  - Trim dead helpers and other dormant surface that still looks active.
+
+- Highest-leverage repo and workflow optimizations:
+  - Continue migrating the oversized legacy test/combinator surface into focused modern tests so it is easier to run, debug, and extend.
+
+## Optimization Candidates
+
+1. Placeholder UI surfaces should either ship or disappear
+- [`src/ffx/help_screen.py`](/home/osgw/.local/src/codex/ffx/src/ffx/help_screen.py) and [`src/ffx/settings_screen.py`](/home/osgw/.local/src/codex/ffx/src/ffx/settings_screen.py) are placeholders.
+- Optimization:
+  - Either remove them from the active UI surface or complete them.
+  - Avoid paying ongoing maintenance cost for unfinished navigation targets.
+- Expected value:
+  - Leaner interface.
+  - Lower UX ambiguity.
+
+2. Several helper functions are unfinished or dead-weight
+- [`src/ffx/helper.py`](/home/osgw/.local/src/codex/ffx/src/ffx/helper.py) contains `permutateList(...): pass`.
+- There are many combinator and conversion placeholders across tests and migrations.
+- Optimization:
+  - Remove dead code, finish it, or isolate it behind a clearly dormant area.
+  - Avoid carrying stubbed utility surface that looks reusable but is not.
+- Expected value:
+  - Smaller mental model.
+  - Less time spent re-evaluating inactive paths.
+
+3. Test suite shape is expensive to understand and likely expensive to run
+- The project still carries a large legacy matrix of combinator files under [`tests/legacy`](/home/osgw/.local/src/codex/ffx/tests/legacy), several placeholder `pass` implementations, and at least one suspicious filename with an embedded space: [`tests/legacy/disposition_combinator_2_3 .py`](/home/osgw/.local/src/codex/ffx/tests/legacy/disposition_combinator_2_3 .py).
+- A first focused replacement slice now exists in [`tests/integration/subtrack_mapping/test_cli_bundle.py`](/home/osgw/.local/src/codex/ffx/tests/integration/subtrack_mapping/test_cli_bundle.py), so the remaining work is migration and consolidation rather than creating the modern test shape from scratch.
+- Optimization:
+  - Continue replacing broad combinator matrices with focused parametrized integration and unit tests.
+  - Retire the bespoke legacy discovery and runner path once equivalent coverage exists.
+  - Normalize file naming and test discovery conventions.
+- Expected value:
+  - Faster contributor onboarding.
+  - Easier CI adoption later.
+
+## Open
+
+- Durable shipped items have been moved into [`README.md`](/home/osgw/.local/src/codex/ffx/README.md) version history through `0.2.6`.
+- Should optimization work focus first on operator-perceived latency, internal maintainability, or correctness-risk cleanup that also has performance upside?
+- Is the long-term supported model still “local Linux workstation plus Textual UI,” or should optimization decisions bias toward a more scriptable/headless CLI?
+
+## Gaps Right Now
+
+- No explicit prioritization owner or milestone for the optimization backlog.
+- No benchmark or timing harness exists for startup, probe, DB, or conversion orchestration overhead.
+- Repo hygiene is still mixed with generated artifacts and some clearly unfinished files.
+- The legacy TMDB-backed `Scenario 4` path is currently blocked by a pattern/track regression: `Patterns must define at least one track before they can be stored.` This surfaced while rerunning TMDB-dependent checks after the zero-track pattern hardening.
+
+## Next
+
+1. Triage the list into quick wins, medium refactors, and long-horizon cleanup.
+2. Tackle the cheapest remaining product-surface cleanup first:
+   - placeholder UI surfaces and dead helper cleanup.
+3. Continue replacing oversized legacy test matrices with focused modern integration and unit coverage.
+4. Triage the legacy `Scenario 4` pattern/track failure and decide whether to fix the harness, adapt it to the zero-track guard, or retire that path during the ongoing test-suite migration.
+
+## Delete When
+
+- Delete this scratchpad once the optimization backlog is either converted into issues/work items or distilled into durable project guidance.
+
+
+
+## TODO: Review styled ASS separate handling

docs/Makefile

@@ -0,0 +1,21 @@
+SPHINXOPTS       ?=
+VENV_SPHINXBUILD = ../.venv/bin/sphinx-build
+SPHINXBUILD      ?= $(if $(wildcard $(VENV_SPHINXBUILD)),$(VENV_SPHINXBUILD),sphinx-build)
+SOURCEDIR        = .
+BUILDDIR         = _build
+
+.PHONY: help clean html linkcheck
+
+help:
+	@echo "Please use 'make <target>' where <target> is one of"
+	@echo "  html       to make standalone HTML files"
+	@echo "  linkcheck  to check all external links for integrity"
+
+clean:
+	rm -rf "$(BUILDDIR)"
+
+html:
+	@$(SPHINXBUILD) -b html "$(SOURCEDIR)" "$(BUILDDIR)/html" $(SPHINXOPTS)
+
+linkcheck:
+	@$(SPHINXBUILD) -b linkcheck "$(SOURCEDIR)" "$(BUILDDIR)/linkcheck" $(SPHINXOPTS)

docs/api.rst

@@ -0,0 +1,31 @@
+API Reference
+=============
+
+This section exposes selected modules that are useful when working on tests,
+diagnostics, process execution, metadata editing, and file probing.
+
+CLI Helpers
+-----------
+
+.. automodule:: ffx.cli
+   :members:
+   :undoc-members:
+
+Process Helpers
+---------------
+
+.. automodule:: ffx.process
+   :members:
+   :undoc-members:
+
+File Probing
+------------
+
+.. automodule:: ffx.file_properties
+
+Metadata Editing
+----------------
+
+.. automodule:: ffx.metadata_editor
+   :members:
+   :undoc-members:

docs/conf.py

@@ -0,0 +1,44 @@
+from __future__ import annotations
+
+from importlib.metadata import PackageNotFoundError, version as package_version
+from pathlib import Path
+import sys
+
+
+ROOT_DIR = Path(__file__).resolve().parents[1]
+SRC_DIR = ROOT_DIR / "src"
+sys.path.insert(0, str(SRC_DIR))
+
+project = "FFX"
+author = "javanaut@maveno.de"
+copyright = "2026, Maveno"
+
+try:
+    release = package_version("ffx")
+except PackageNotFoundError:
+    release = "0.0.0"
+version = release
+
+extensions = [
+    "sphinx.ext.autodoc",
+    "sphinx.ext.napoleon",
+    "sphinx.ext.viewcode",
+    "sphinx_copybutton",
+]
+
+source_suffix = {
+    ".rst": "restructuredtext",
+}
+
+templates_path = ["_templates"]
+exclude_patterns = ["_build", "Thumbs.db", ".DS_Store"]
+
+html_theme = "sphinx_rtd_theme"
+html_title = "FFX"
+html_static_path = []
+
+autodoc_typehints = "description"
+autodoc_member_order = "bysource"
+napoleon_google_docstring = True
+napoleon_numpy_docstring = True
+

docs/development.rst

@@ -0,0 +1,50 @@
+Development
+===========
+
+The repo-local ``.venv`` is the preferred environment for contributors working
+on tests or documentation:
+
+.. code-block:: sh
+
+   tests/prepare.sh
+
+The preparation script installs the package in editable mode with both test and
+documentation extras:
+
+.. code-block:: text
+
+   .[test,docs]
+
+Run Tests
+---------
+
+Run the modern pytest suite:
+
+.. code-block:: sh
+
+   .venv/bin/python -m pytest --ignore=tests/legacy --ignore=tests/support tests
+
+The legacy harness remains available separately and is intentionally not part of
+the default pytest run.
+
+Build Docs
+----------
+
+Build HTML documentation:
+
+.. code-block:: sh
+
+   .venv/bin/sphinx-build -b html docs docs/_build/html
+
+The same command is wrapped by the Sphinx ``Makefile``:
+
+.. code-block:: sh
+
+   make -C docs html
+
+VS Code
+-------
+
+The repository includes ``.vscode/extensions.json`` with recommended
+extensions, including Esbonio for Sphinx language-server support. The workspace
+settings point Python tooling and Esbonio at the repo-local ``.venv``.

docs/file_formats.md

@@ -1,170 +0,0 @@
-# File Formats
-
-This document captures source-file-format notes that complement the normative
-requirements in `requirements/source_file_formats.md`.
-
-The first documented format is a Matroska source that carries styled ASS/SSA
-subtitle streams together with embedded font attachments.
-
-## Styled ASS In Matroska With Embedded Fonts
-
-These files are typically `.mkv` releases where subtitle rendering quality
-depends on keeping both parts of the subtitle package together:
-
-- one or more subtitle streams with codec `ass`
-- one or more attachment streams that embed font files used by those subtitles
-
-This matters because ASS subtitles are not plain text subtitles in the narrow
-WebVTT sense. They can carry layout, styling, positioning, karaoke, signs, and
-other typesetting effects. If the matching embedded fonts are lost, consumers
-can still see subtitle text but the intended styling and sometimes glyph
-coverage can be degraded.
-
-For FFX this format is special because the ASS subtitle streams should remain
-normally editable and mappable, while the related font attachments should be
-transported unchanged.
-
-## Observed Sample
-
-Assessment date: `2026-04-17`
-
-Observed sample file:
-
-- `tests/assets/boruto_s01e283_ssa.mkv`
-
-Commands used for assessment:
-
-```bash
-ffprobe tests/assets/boruto_s01e283_ssa.mkv
-ffprobe -hide_banner -show_format -show_streams -of json tests/assets/boruto_s01e283_ssa.mkv
-```
-
-Observed stream layout:
-
-| Stream index | Kind | Key details |
-| --- | --- | --- |
-| `0` | video | `codec_name=h264` |
-| `1` | audio | `codec_name=aac`, `language=jpn` |
-| `2` | subtitle | `codec_name=ass`, `language=ger`, default |
-| `3` | subtitle | `codec_name=ass`, `language=eng` |
-| `4`-`13` | attachment | `tags.mimetype=font/ttf`, `.ttf` filenames |
-
-Observed attachment filenames:
-
-- `AmazonEmberTanuki-Italic.ttf`
-- `AmazonEmberTanuki-Regular.ttf`
-- `Arial.ttf`
-- `Arial Bold.ttf`
-- `Georgia.ttf`
-- `Times New Roman.ttf`
-- `Times New Roman Bold.ttf`
-- `Trebuchet MS.ttf`
-- `Verdana.ttf`
-- `Verdana Bold.ttf`
-
-Important probe behavior from the real sample:
-
-- Plain `ffprobe` lists the font streams as `Attachment: none`.
-- Plain `ffprobe` also prints warnings such as `Could not find codec
-  parameters for stream 4 (Attachment: none): unknown codec` and later
-  `Unsupported codec with id 0 for input stream ...`.
-- The JSON produced by `FileProperties.FFPROBE_COMMAND_TOKENS`
-  (`ffprobe -hide_banner -show_format -show_streams -of json`) still exposes
-  the attachment streams clearly through `codec_type="attachment"` and the
-  attachment tags.
-- In that JSON, the attachment streams do not expose `codec_name`.
-
-This last point is important for FFX: robust detection must not depend on
-attachment `codec_name` being present.
-
-## Detection Guidance
-
-Current known indicators for this format are:
-
-- one or more subtitle streams with `codec_type="subtitle"` and
-  `codec_name="ass"`
-- one or more attachment streams with `codec_type="attachment"`
-- attachment tags that identify embedded fonts, especially
-  `tags.mimetype="font/ttf"`
-- attachment filenames that end in `.ttf`
-
-The pattern can vary. FFX should therefore treat the above as a cluster of
-signals rather than an exact signature tied to one file.
-
-Inference from the observed sample plus FFmpeg documentation:
-
-- MIME matching should not be limited to `font/ttf` alone.
-- The Boruto sample uses `font/ttf`.
-- FFmpeg's Matroska attachment example uses
-  `mimetype=application/x-truetype-font` for a `.ttf` attachment.
-- Detection should therefore normalize multiple TTF-like MIME values rather
-  than depend on a single exact string.
-
-## Processing Expectations In FFX
-
-The format-specific requirements live in
-`requirements/source_file_formats.md`. In practical terms, FFX should:
-
-- recognize the ASS-plus-font-attachment pattern even when attachment probe
-  data is incomplete
-- tell the operator that the pattern was detected and that special handling is
-  being used
-- reject sidecar subtitle import for such sources, because converting or
-  replacing these subtitle tracks with ordinary external text subtitles would
-  break the intended subtitle package
-- continue to allow normal manipulation of the ASS subtitle tracks themselves
-- preserve the font attachment streams unchanged
-
-## FFmpeg Notes
-
-Relevant FFmpeg documentation confirms several behaviors that line up with
-FFX's needs:
-
-- FFmpeg documents `-attach` as adding an attachment stream to the output, and
-  explicitly names Matroska fonts used in subtitle rendering as an example.
-- FFmpeg documents attachment streams as regular streams that are created after
-  the mapped media streams.
-- FFmpeg documents `-dump_attachment` for extracting attachment streams, which
-  is useful for debugging or validating a source file's embedded fonts.
-- FFmpeg's Matroska example requires a `mimetype` metadata tag for attached
-  fonts, which is consistent with using attachment tags as detection signals.
-- FFmpeg also notes that attachments are implemented as codec extradata. That
-  helps explain why probe output for attachment streams can look different from
-  ordinary audio, video, and subtitle streams.
-
-Implication for FFX:
-
-- Attachment preservation is not an optional cosmetic feature for this format.
-  It is part of preserving the subtitle package correctly.
-
-## Jellyfin Notes
-
-Jellyfin's documentation also supports keeping this format intact:
-
-- Jellyfin's subtitle compatibility table lists `ASS/SSA` as supported in
-  `MKV` and not supported in `MP4`.
-- Jellyfin notes that when subtitles must be transcoded, they are either
-  converted to a supported format or burned into the video, and burning them in
-  is the most CPU-intensive path.
-- Jellyfin's subtitle-extraction example for `SSA/ASS` first dumps attachment
-  streams and then extracts the ASS subtitle stream, which reflects the real
-  relationship between ASS subtitles and embedded fonts in MKV releases.
-- Jellyfin's font documentation says text-based subtitles require fonts to
-  render properly.
-- Jellyfin's configuration documentation says the web client uses configured
-  fallback fonts for ASS subtitles when other fonts such as MKV attachments or
-  client-side fonts are not available.
-
-Inference from the Jellyfin compatibility tables:
-
-- Keeping this subtitle format in Matroska is the safest interoperability
-  choice for Jellyfin consumers.
-- Converting the subtitle payload to WebVTT would lose styled ASS behavior.
-- Dropping the attachment streams would force client or fallback font
-  substitution and can change appearance or glyph coverage.
-
-## References
-
-- FFmpeg documentation: https://ffmpeg.org/ffmpeg.html
-- Jellyfin codec support: https://jellyfin.org/docs/general/clients/codec-support/
-- Jellyfin configuration and fonts: https://jellyfin.org/docs/general/administration/configuration/

docs/file_formats.rst

@@ -0,0 +1,192 @@
+File Formats
+============
+
+This document captures source-file-format notes that complement the normative
+requirements in ``requirements/source_file_formats.md``.
+
+The first documented format is a Matroska source that carries styled ASS/SSA
+subtitle streams together with embedded font attachments.
+
+Styled ASS In Matroska With Embedded Fonts
+------------------------------------------
+
+These files are typically ``.mkv`` releases where subtitle rendering quality
+depends on keeping both parts of the subtitle package together:
+
+* one or more subtitle streams with codec ``ass``
+* one or more attachment streams that embed font files used by those subtitles
+
+This matters because ASS subtitles are not plain text subtitles in the narrow
+WebVTT sense. They can carry layout, styling, positioning, karaoke, signs, and
+other typesetting effects. If the matching embedded fonts are lost, consumers
+can still see subtitle text but the intended styling and sometimes glyph
+coverage can be degraded.
+
+For FFX this format is special because the ASS subtitle streams should remain
+normally editable and mappable, while the related font attachments should be
+transported unchanged.
+
+Observed Sample
+---------------
+
+Assessment date: ``2026-04-17``
+
+Observed sample file:
+
+* ``tests/assets/boruto_s01e283_ssa.mkv``
+
+Commands used for assessment:
+
+.. code-block:: bash
+
+   ffprobe tests/assets/boruto_s01e283_ssa.mkv
+   ffprobe -hide_banner -show_format -show_streams -of json tests/assets/boruto_s01e283_ssa.mkv
+
+Observed stream layout:
+
+.. list-table::
+   :header-rows: 1
+
+   * - Stream index
+     - Kind
+     - Key details
+   * - ``0``
+     - video
+     - ``codec_name=h264``
+   * - ``1``
+     - audio
+     - ``codec_name=aac``, ``language=jpn``
+   * - ``2``
+     - subtitle
+     - ``codec_name=ass``, ``language=ger``, default
+   * - ``3``
+     - subtitle
+     - ``codec_name=ass``, ``language=eng``
+   * - ``4``-``13``
+     - attachment
+     - ``tags.mimetype=font/ttf``, ``.ttf`` filenames
+
+Observed attachment filenames:
+
+* ``AmazonEmberTanuki-Italic.ttf``
+* ``AmazonEmberTanuki-Regular.ttf``
+* ``Arial.ttf``
+* ``Arial Bold.ttf``
+* ``Georgia.ttf``
+* ``Times New Roman.ttf``
+* ``Times New Roman Bold.ttf``
+* ``Trebuchet MS.ttf``
+* ``Verdana.ttf``
+* ``Verdana Bold.ttf``
+
+Important probe behavior from the real sample:
+
+* Plain ``ffprobe`` lists the font streams as ``Attachment: none``.
+* Plain ``ffprobe`` also prints warnings such as ``Could not find codec
+  parameters for stream 4 (Attachment: none): unknown codec`` and later
+  ``Unsupported codec with id 0 for input stream ...``.
+* The JSON produced by ``FileProperties.FFPROBE_COMMAND_TOKENS``
+  (``ffprobe -hide_banner -show_format -show_streams -of json``) still exposes
+  the attachment streams clearly through ``codec_type="attachment"`` and the
+  attachment tags.
+* In that JSON, the attachment streams do not expose ``codec_name``.
+
+This last point is important for FFX: robust detection must not depend on
+attachment ``codec_name`` being present.
+
+Detection Guidance
+------------------
+
+Current known indicators for this format are:
+
+* one or more subtitle streams with ``codec_type="subtitle"`` and
+  ``codec_name="ass"``
+* one or more attachment streams with ``codec_type="attachment"``
+* attachment tags that identify embedded fonts, especially
+  ``tags.mimetype="font/ttf"``
+* attachment filenames that end in ``.ttf``
+
+The pattern can vary. FFX should therefore treat the above as a cluster of
+signals rather than an exact signature tied to one file.
+
+Inference from the observed sample plus FFmpeg documentation:
+
+* MIME matching should not be limited to ``font/ttf`` alone.
+* The Boruto sample uses ``font/ttf``.
+* FFmpeg's Matroska attachment example uses
+  ``mimetype=application/x-truetype-font`` for a ``.ttf`` attachment.
+* Detection should therefore normalize multiple TTF-like MIME values rather
+  than depend on a single exact string.
+
+Processing Expectations In FFX
+------------------------------
+
+The format-specific requirements live in
+``requirements/source_file_formats.md``. In practical terms, FFX should:
+
+* recognize the ASS-plus-font-attachment pattern even when attachment probe data
+  is incomplete
+* tell the operator that the pattern was detected and that special handling is
+  being used
+* reject sidecar subtitle import for such sources, because converting or
+  replacing these subtitle tracks with ordinary external text subtitles would
+  break the intended subtitle package
+* continue to allow normal manipulation of the ASS subtitle tracks themselves
+* preserve the font attachment streams unchanged
+
+FFmpeg Notes
+------------
+
+Relevant FFmpeg documentation confirms several behaviors that line up with
+FFX's needs:
+
+* FFmpeg documents ``-attach`` as adding an attachment stream to the output, and
+  explicitly names Matroska fonts used in subtitle rendering as an example.
+* FFmpeg documents attachment streams as regular streams that are created after
+  the mapped media streams.
+* FFmpeg documents ``-dump_attachment`` for extracting attachment streams, which
+  is useful for debugging or validating a source file's embedded fonts.
+* FFmpeg's Matroska example requires a ``mimetype`` metadata tag for attached
+  fonts, which is consistent with using attachment tags as detection signals.
+* FFmpeg also notes that attachments are implemented as codec extradata. That
+  helps explain why probe output for attachment streams can look different from
+  ordinary audio, video, and subtitle streams.
+
+Implication for FFX:
+
+* Attachment preservation is not an optional cosmetic feature for this format.
+  It is part of preserving the subtitle package correctly.
+
+Jellyfin Notes
+--------------
+
+Jellyfin's documentation also supports keeping this format intact:
+
+* Jellyfin's subtitle compatibility table lists ``ASS/SSA`` as supported in
+  ``MKV`` and not supported in ``MP4``.
+* Jellyfin notes that when subtitles must be transcoded, they are either
+  converted to a supported format or burned into the video, and burning them in
+  is the most CPU-intensive path.
+* Jellyfin's subtitle-extraction example for ``SSA/ASS`` first dumps attachment
+  streams and then extracts the ASS subtitle stream, which reflects the real
+  relationship between ASS subtitles and embedded fonts in MKV releases.
+* Jellyfin's font documentation says text-based subtitles require fonts to
+  render properly.
+* Jellyfin's configuration documentation says the web client uses configured
+  fallback fonts for ASS subtitles when other fonts such as MKV attachments or
+  client-side fonts are not available.
+
+Inference from the Jellyfin compatibility tables:
+
+* Keeping this subtitle format in Matroska is the safest interoperability choice
+  for Jellyfin consumers.
+* Converting the subtitle payload to WebVTT would lose styled ASS behavior.
+* Dropping the attachment streams would force client or fallback font
+  substitution and can change appearance or glyph coverage.
+
+References
+----------
+
+* FFmpeg documentation: https://ffmpeg.org/ffmpeg.html
+* Jellyfin codec support: https://jellyfin.org/docs/general/clients/codec-support/
+* Jellyfin configuration and fonts: https://jellyfin.org/docs/general/administration/configuration/

docs/index.rst

@@ -0,0 +1,25 @@
+FFX Documentation
+=================
+
+FFX is a local command-line and Textual terminal UI for inspecting TV episode
+files, storing normalization rules, and converting media into predictable
+archive-ready outputs.
+
+This documentation covers operator setup, day-to-day command usage, contributor
+workflow, format-specific notes, and generated API references for the smaller
+utility modules.
+
+.. toctree::
+   :maxdepth: 2
+   :caption: User Guide
+
+   installation
+   usage
+   file_formats
+
+.. toctree::
+   :maxdepth: 2
+   :caption: Contributor Guide
+
+   development
+   api

docs/installation.rst

@@ -0,0 +1,52 @@
+Installation
+============
+
+FFX is designed for a Linux-like workstation with local command execution. The
+runtime media tools must be available on ``PATH``:
+
+* ``ffmpeg``
+* ``ffprobe``
+* ``cpulimit``
+
+User Bundle
+-----------
+
+The persistent user installation is prepared with the two-step flow described in
+the project README:
+
+.. code-block:: sh
+
+   bash tools/setup.sh
+   bash tools/configure_workstation.sh
+
+``tools/setup.sh`` creates the long-lived bundle virtualenv at
+``~/.local/share/ffx.venv`` and exposes the ``ffx`` command. The workstation
+script checks system tools and seeds local config directories.
+
+Local Test And Docs Environment
+-------------------------------
+
+Contributor test and documentation work uses the repo-local virtualenv:
+
+.. code-block:: sh
+
+   tests/prepare.sh
+
+The script creates ``.venv``, installs FFX in editable mode with test and docs
+extras, and verifies the Sphinx toolchain. Use check-only mode when you only
+want to inspect readiness:
+
+.. code-block:: sh
+
+   tests/prepare.sh --check
+
+Documentation Build
+-------------------
+
+After preparation, build the documentation with:
+
+.. code-block:: sh
+
+   .venv/bin/sphinx-build -b html docs docs/_build/html
+
+The generated site starts at ``docs/_build/html/index.html``.

docs/make.bat

@@ -0,0 +1,42 @@
+@ECHO OFF
+
+pushd %~dp0
+
+if "%SPHINXBUILD%" == "" if exist ..\.venv\Scripts\sphinx-build.exe (
+	set SPHINXBUILD=..\.venv\Scripts\sphinx-build.exe
+)
+if "%SPHINXBUILD%" == "" set SPHINXBUILD=sphinx-build
+set SOURCEDIR=.
+set BUILDDIR=_build
+
+%SPHINXBUILD% >NUL 2>NUL
+if errorlevel 9009 (
+	echo.
+	echo The 'sphinx-build' command was not found. Make sure Sphinx is installed,
+	echo then set SPHINXBUILD to the full path if needed.
+	exit /b 1
+)
+
+if "%1" == "" goto help
+if "%1" == "html" goto html
+if "%1" == "linkcheck" goto linkcheck
+echo.
+echo Unknown target "%1".
+goto help
+
+:html
+%SPHINXBUILD% -b html %SOURCEDIR% %BUILDDIR%\html %SPHINXOPTS%
+goto end
+
+:linkcheck
+%SPHINXBUILD% -b linkcheck %SOURCEDIR% %BUILDDIR%\linkcheck %SPHINXOPTS%
+goto end
+
+:help
+echo.
+echo Please use 'make.bat ^<target^>' where ^<target^> is one of
+echo   html       to make standalone HTML files
+echo   linkcheck  to check all external links for integrity
+
+:end
+popd

docs/usage.rst

@@ -0,0 +1,97 @@
+Usage
+=====
+
+FFX exposes a single ``ffx`` command with subcommands for inspection,
+conversion, metadata editing, setup, and maintenance.
+
+Inspect Files
+-------------
+
+Open the inspection workflow for one or more files:
+
+.. code-block:: sh
+
+   ffx inspect /path/to/episode.mkv
+
+Print resolved season-shift mappings without opening the TUI:
+
+.. code-block:: sh
+
+   ffx inspect --shift /path/to/episode.mkv
+
+Convert Files
+-------------
+
+Convert one or more source files using stored rules where available:
+
+.. code-block:: sh
+
+   ffx convert /path/to/episode.mkv
+
+Useful overrides include:
+
+* ``--no-pattern`` to skip database pattern matching
+* ``--show``, ``--season``, and ``--episode`` for explicit episode identity
+* ``--output-directory`` for generated output placement
+* ``--subtitle-directory`` for source-basename sidecar subtitle imports
+* ``--subtitle-prefix`` for explicit or configured-prefix subtitle imports
+* ``--subtitle-extension`` to select the imported sidecar format (default:
+  ``vtt``)
+* ``--yes`` to accept a valid partial sidecar set without prompting
+* ``--copy-video`` or ``--copy-audio`` to preserve selected stream types
+* ``--rename-only`` for filename normalization without media rewriting
+
+Directory-only subtitle import matches the source basename. For example,
+``A2_t01.mkv`` discovers files such as ``A2_t01_2_deu_DEF.vtt`` in the
+provided directory:
+
+.. code-block:: sh
+
+   ffx convert --subtitle-directory /path/to/subtitles A2_t01.mkv
+
+Select a different sidecar set by extension, with or without the leading dot:
+
+.. code-block:: sh
+
+   ffx convert --subtitle-directory /path/to/subtitles \
+     --subtitle-extension .mkv A2_t01.mkv
+
+When only some source subtitle tracks have matching sidecar files, conversion
+asks for confirmation. Use ``--yes`` to substitute that valid subset without
+prompting. ``--yes`` also permits this case when ``--no-prompt`` is set.
+
+Manage Shows And Patterns
+-------------------------
+
+Open the Textual interface for show and pattern management:
+
+.. code-block:: sh
+
+   ffx shows
+
+Extract Streams
+---------------
+
+Extract streams from a file:
+
+.. code-block:: sh
+
+   ffx unmux /path/to/episode.mkv
+
+For subtitle-only extraction:
+
+.. code-block:: sh
+
+   ffx unmux --subtitles-only --label show-name /path/to/episode.mkv
+
+Detect Crop
+-----------
+
+Ask FFmpeg to suggest crop parameters:
+
+.. code-block:: sh
+
+   ffx cropdetect /path/to/episode.mkv
+
+The default sampling window is controlled by the application defaults and can be
+overridden with command options.

guidance/workflow/optional/lean-interface-iteration.md

@@ -0,0 +1,28 @@
+# Lean Interface Iteration
+
+Rule set name: `lean-interface-iteration`
+
+Rule set ID: `LII`
+
+Status: optional, prompt-activated only
+
+Trigger examples:
+
+- `Apply the lean-interface-iteration rules.`
+- `Apply LII rules.`
+
+LII-0001: Apply this rule set only when it is explicitly requested in the prompt.
+
+LII-0002: The target of work under this rule set is the iterated product state for the addressed iteration only.
+
+LII-0003: Optimize the addressed interface toward the leanest and least complex model that still satisfies the iteration order.
+
+LII-0004: Backward compatibility, legacy aliases, and compatibility shims are not required unless the prompt explicitly asks to preserve them.
+
+LII-0005: Prefer one authoritative interface over multiple overlapping parameters, flags, or naming variants.
+
+LII-0006: Remove or avoid transitional interface layers when they are not required by the addressed iteration order.
+
+LII-0007: Update affected tests, guidance, requirements, and documentation so they describe the simplified interface model rather than a mixed legacy-and-new model.
+
+LII-0008: Never change behavior, interfaces, or surrounding areas that are not addressed by the current iteration order.

guidance/workflow/optional/preparation-script-design.md

@@ -0,0 +1,56 @@
+# Preparation Script Design
+
+Rule set name: `preparation-script-design`
+
+Rule set ID: `PSD`
+
+Status: optional, prompt-activated only
+
+Trigger examples:
+
+- `Apply the preparation-script-design rules.`
+- `Apply PSD rules.`
+
+PSD-0001: Apply this rule set only when it is explicitly requested in the prompt.
+
+PSD-0002: Use this rule set for scripts whose purpose is to prepare, verify, or expose a local development or automation environment rather than to perform product runtime behavior.
+
+PSD-0003: Keep a preparation script focused on environment readiness, dependency installation, local helper exposure, and clear verification output; do not mix unrelated product logic into the script.
+
+PSD-0004: Design the script to be idempotent so repeated runs converge on the same prepared state without unnecessary reinstallation or destructive side effects.
+
+PSD-0005: Provide a verification-only mode such as `--check` that reports readiness without installing, modifying, or creating dependencies.
+
+PSD-0006: Separate component checks from installation steps so the script can report what is missing before or after attempted remediation.
+
+PSD-0007: Group required capabilities into clear purpose-oriented sections such as support toolchains, local package bundles, generated environment helpers, or other relevant readiness areas instead of presenting one undifferentiated dependency list.
+
+PSD-0008: Prefer explicit per-component check helpers over opaque one-shot checks so failures remain traceable and easy to extend.
+
+PSD-0009: Generate or update environment helper files only when they provide a stable, reusable way to expose repo-local or workspace-local tools, paths, or environment variables.
+
+PSD-0010: Generated environment helper files shall be safe to source multiple times and should avoid duplicating path entries or clobbering unrelated user environment state.
+
+PSD-0011: When a preparation flow seeds optional user-owned files such as config templates, do so non-destructively by creating them only when absent unless the prompt explicitly requests overwrite behavior.
+
+PSD-0012: Report status in a concise scan-friendly line format of the shape `[status] Label: detail`, where the label names the checked component and the detail string stays short and specific.
+
+PSD-0013: Prefer a small canonical status vocabulary in those report lines, with `ok` for satisfied checks, `warn` for non-blocking gaps, and a failure status such as `failed` for blocking or unsuccessful states.
+
+PSD-0014: When a preparation script uses terminal colors in its status output, apply a consistent severity mapping so `ok` is green, `warn` is yellow, and all other status levels are red.
+
+PSD-0015: In bracketed status markers such as `[ok]` or `[warn]`, keep the square brackets uncolored and apply the severity color only to the inner status text.
+
+PSD-0016: Colorized status output shall degrade safely in non-terminal or non-color contexts so the script remains readable and automation-friendly without ANSI support.
+
+PSD-0017: End with an explicit readiness conclusion that distinguishes between successful preparation, incomplete prerequisites, and failed installation attempts.
+
+PSD-0018: Installation logic should use the narrowest supported platform-specific package-manager actions necessary for the declared scope and should fail clearly when no supported installation path is available.
+
+PSD-0019: Treat repo-local helper tooling and local package installation boundaries explicitly rather than assuming global installs, especially when the prepared environment is intended to be reproducible.
+
+PSD-0020: Keep the script suitable for both interactive local developer use and non-interactive automation checks by avoiding prompts during normal execution unless the prompt explicitly requires interactivity.
+
+PSD-0021: When a script depends on generated helper files or adjacent validation helpers, update those supporting files only as needed to keep the preparation flow coherent and usable.
+
+PSD-0022: Verify shell syntax after changes and, when feasible, run a dry readiness check so the resulting preparation flow is validated rather than only written.

pyproject.toml

@@ -31,6 +31,12 @@ Issues = "https://gitea.maveno.de/Javanaut/ffx/issues"
 test = [
   "pytest",
 ]
+docs = [
+  "esbonio",
+  "sphinx",
+  "sphinx-copybutton",
+  "sphinx-rtd-theme",
+]
 
 [build-system]
 requires = [

requirements/architecture.md

@@ -0,0 +1,98 @@
+# Architecture
+
+## Architecture Goals
+
+- Keep the tool small, local, and easy to reason about.
+- Separate media inspection, stored normalization rules, and conversion execution clearly enough that users can inspect and adjust behavior.
+- Favor explicit local state and deterministic rule application over opaque automation.
+- Make external runtime dependencies and platform assumptions visible.
+
+## System Context
+
+- Primary actors:
+  - Local operator running the CLI.
+  - Local operator using the Textual TUI to inspect files and maintain rules.
+- External systems:
+  - `ffprobe` for media introspection.
+  - `ffmpeg` for conversion and extraction.
+  - TMDB API for optional show and episode metadata.
+  - Local filesystem for source media, generated outputs, subtitles, logs, config, and database files.
+- Data entering the system:
+  - Media container and stream metadata from source files.
+  - Regex patterns and per-show normalization rules entered in the TUI.
+  - Optional config values from `~/.local/etc/ffx.json`.
+  - Optional TMDB identifiers and CLI overrides.
+  - Optional external subtitle files.
+- Data leaving the system:
+  - Normalized output media files.
+  - Extracted stream files from unmux operations.
+  - SQLite rows representing shows, patterns, tracks, tags, shifted seasons, and properties.
+  - Local log output and console messages.
+
+## High-Level Building Blocks
+
+- Frontend, CLI, API, or worker:
+  - A Click-based CLI in [`src/ffx/cli.py`](/home/osgw/.local/src/codex/ffx/src/ffx/cli.py), exposed as the `ffx` command and via `python -m ffx`, including lightweight maintenance wrappers for bundle setup, workstation preparation, and upgrade tasks.
+  - A Textual terminal UI rooted in [`src/ffx/ffx_app.py`](/home/osgw/.local/src/codex/ffx/src/ffx/ffx_app.py) with screens for shows, patterns, file inspection, tracks, tags, and shifted seasons.
+- Core business logic:
+  - Descriptor objects model media files, shows, and tracks.
+  - Controllers encapsulate CRUD operations and workflow orchestration for shows, patterns, tags, tracks, season shifts, configuration, and conversion.
+  - `MediaDescriptorChangeSet` computes differences between a file and its stored target schema to drive metadata and disposition updates.
+  - File inspection caches combined `ffprobe` data and crop-detection results per source and sampling window within one process to avoid repeated subprocess work.
+- Storage:
+  - SQLite via SQLAlchemy ORM, with schema rooted in shows, patterns, tracks, media tags, track tags, shifted seasons, and generic properties.
+  - Ordered schema migrations are loaded dynamically from per-version-step modules under [`src/ffx/model/migration/`](/home/osgw/.local/src/codex/ffx/src/ffx/model/migration/).
+  - A configuration JSON file supplies optional path, metadata-filtering, and filename-template settings.
+- Integration adapters:
+  - Process execution wrapper for `ffmpeg`, `ffprobe`, `nice`, and `cpulimit`, with explicit disabled states for niceness and CPU limiting, support for both absolute `cpulimit` values and machine-wide percent input, and a combined `cpulimit -- nice -n ... <command>` execution shape when both limits are configured.
+  - HTTP adapter for TMDB via `requests`.
+
+## Data And Interface Notes
+
+- Key entities or records:
+  - `Show`: canonical TV show metadata plus digit-formatting rules, optional show-level notes, and an optional show-level encoding-quality fallback.
+  - `Pattern`: regex rule tying filenames to one show and one target media schema.
+  - `Track` and `TrackTag`: persisted target stream records, codec, dispositions, audio layout, and stream-level tags. Detailed source-to-target mapping rules live in `requirements/subtrack_mapping.md`.
+  - `MediaTag`: persisted container-level metadata for a pattern.
+  - `ShiftedSeason`: mapping from source numbering ranges to adjusted season and episode numbers, owned either by a show as fallback or by a pattern as override.
+  - `Property`: internal key-value storage currently used for database versioning.
+- External interfaces:
+  - CLI commands for conversion, inspection, extraction, and crop detection.
+  - TUI workflows for rule authoring and rule maintenance.
+  - Environment variable `TMDB_API_KEY` for TMDB access.
+  - Config keys `databasePath`, `logDirectory`, and `outputFilenameTemplate`, plus optional metadata-filter rules.
+- Validation rules:
+  - Only supported media-file extensions are accepted for conversion.
+  - Stored database version must either match the runtime-required version already or have a supported sequential migration path to it.
+  - A normalized descriptor may have at most one default and one forced stream per relevant track type.
+  - Shifted-season ranges are intended not to overlap within the same owner scope and season, and runtime resolution prefers pattern-owned matches over show-owned matches.
+  - TMDB lookups require a show ID and season and episode numbers.
+- Error-handling approach:
+  - User-facing operational failures are raised as `click.ClickException` or warnings.
+  - Ambiguous default and forced stream states trigger prompts unless `--no-prompt` is set, in which case the command fails fast.
+  - External-process failures and invalid media are surfaced through logs and command errors rather than retries, except for TMDB rate-limit retries.
+
+## Deployment And Operations
+
+- Runtime environment:
+  - Local Python environment with the package installed and `ffmpeg`, `ffprobe`, `nice`, and `cpulimit` available on `PATH`.
+- Deployment shape:
+  - Single-process command execution on demand; no daemon, queue, or network service of its own.
+- Secrets and configuration handling:
+  - TMDB secret is read from `TMDB_API_KEY`.
+  - User config is read from `~/.local/etc/ffx.json`.
+  - Database path may also be overridden per command via `--database-file`.
+- Logging and monitoring approach:
+  - File and console logging configured per invocation.
+  - Default log file path is `~/.local/var/log/ffx.log`.
+  - No dedicated monitoring integration is present.
+
+## Open Technical Questions
+
+- Question: Should Linux-specific assumptions such as `/dev/null`, `nice`, `cpulimit`, and `~/.local` remain part of the supported-platform contract?
+- Risk: Portability and operational behavior are underspecified for non-Linux environments.
+- Next decision needed: Either document Linux-like systems as the official support boundary or refactor the process and path handling for broader portability.
+
+- Question: Should placeholder TUI surfaces such as settings and help become part of the required product surface or stay explicitly out of scope?
+- Risk: The UI appears broader than the actually finished feature set.
+- Next decision needed: Either remove or complete placeholder screens and update requirements accordingly.

requirements/metadata_editor.md

@@ -0,0 +1,211 @@
+# Metadata Editor
+
+This file defines the requirements for a database-free interactive metadata
+editor command derived from the current file-inspection UI.
+
+Feasibility from the current codebase: yes, with a moderate refactor.
+
+The strongest reusable pieces already exist:
+
+- `ffprobe`-backed media probing through `FileProperties` and `MediaDescriptor`
+- descriptor-level metadata and disposition mutation through `MediaDescriptor`
+  and `TrackDescriptor`
+- diff and ffmpeg token generation through `MediaDescriptorChangeSet`
+- stream-copy remux execution through `FfxController` with `VideoEncoder.COPY`
+- reusable tag and track edit dialogs in the Textual UI
+
+The main missing pieces are:
+
+- a CLI bootstrap path that does not initialize SQLite
+- a probe-only path that does not instantiate database-backed controllers
+- a clean separation between original file state and editable draft state
+- a safe temporary-output and replace workflow for writing changes back to the
+  same file path
+
+## Scope
+
+- One new command: `ffx edit <file>`
+- One-file interactive editing through a Textual screen derived from
+  `MediaDetailsScreen`
+- Editing container-level metadata and per-stream metadata already visible in
+  the application
+- Editing stream dispositions that are represented as metadata-like output
+  state, especially `default` and `forced`
+- Writing the result back to the original file path through a temporary output
+  file and replace step
+
+## Out Of Scope
+
+- SQLite reads, writes, migrations, or pattern matching
+- TMDB lookups, show selection, pattern selection, or shifted-season logic
+- Batch editing multiple files in one command invocation
+- Video or audio transcoding
+- Container changes, filename changes, or rename workflows
+- Stream add, stream delete, stream reorder, or stream substitution from
+  external files in the first release
+- Editing technical stream identity such as codec, stream type, source index,
+  or audio layout in the first release
+- Chapter editing
+
+## Terms
+
+- `baseline descriptor`: immutable in-memory representation of the file as last
+  probed from disk
+- `draft descriptor`: mutable in-memory representation of the desired output
+  state
+- `edit mode`: the database-free TUI mode used by `ffx edit`
+- `planned changes`: user-visible summary of the differences between baseline
+  and draft plus any configured cleanup actions
+- `temporary output file`: the write target used before replacing the original
+  file path
+
+## Rules
+
+- `METADATA_EDITOR-0001`: The system shall provide a command `ffx edit <file>`
+  that requires exactly one existing media file path and opens an interactive
+  Textual editor for that file.
+- `METADATA_EDITOR-0002`: `ffx edit` shall not initialize SQLite, shall not
+  open the configured database file, shall not prompt for database migration,
+  and shall not instantiate any controller that depends on `context['database']`.
+- `METADATA_EDITOR-0003`: `ffx edit` may still read configuration and logging
+  settings from `~/.local/etc/ffx.json`, but any global database option shall
+  have no effect on this command's behavior.
+- `METADATA_EDITOR-0004`: Edit mode shall be derived from the current
+  `MediaDetailsScreen` behavior and layout where practical, but all DB-only UI
+  elements and actions such as show selection, pattern input, and pattern CRUD
+  actions shall be hidden, disabled, or replaced.
+- `METADATA_EDITOR-0005`: Edit mode shall keep the baseline descriptor and the
+  draft descriptor as separate objects. Editing actions shall mutate only the
+  draft descriptor until the operator explicitly applies changes.
+- `METADATA_EDITOR-0006`: The application shall keep raw metadata values
+  separate from rendered labels. Rich or Textual markup may be used for
+  presentation, but it shall never be stored in descriptor state, reused as
+  source data, or written into the media file.
+- `METADATA_EDITOR-0007`: The planned-changes view shall compare the baseline
+  descriptor with the draft descriptor using `MediaDescriptorChangeSet` or an
+  equivalent descriptor-diff mechanism. It shall no longer mean `file -> db`.
+- `METADATA_EDITOR-0008`: The editor shall support container-tag add, edit, and
+  delete operations on the draft descriptor.
+- `METADATA_EDITOR-0009`: The editor shall support per-stream metadata edit
+  operations on the draft descriptor, including at least language, title, and
+  arbitrary stream tag key-value pairs.
+- `METADATA_EDITOR-0010`: The editor shall support setting and clearing
+  `default` and `forced` dispositions in the draft descriptor, while enforcing
+  that there is at most one `default` and at most one `forced` stream per track
+  type.
+- `METADATA_EDITOR-0011`: The first released editor scope shall treat technical
+  stream structure as immutable. A user shall not be able to change stream
+  count, output order, codec, track type, audio layout, or source-index
+  mapping through `ffx edit`.
+- `METADATA_EDITOR-0012`: The track-edit UI used in edit mode shall therefore
+  expose only metadata fields and supported disposition fields. Structural
+  fields that are editable in pattern-authoring workflows shall be read-only or
+  absent in edit mode.
+- `METADATA_EDITOR-0013`: The command shall write changes through an ffmpeg
+  stream-copy remux workflow only. No transcoding shall be performed as part of
+  `ffx edit`.
+- `METADATA_EDITOR-0013A`: The ffmpeg invocation used by `ffx edit` shall map
+  all source streams with `-map 0` and shall copy all mapped streams with a
+  single `-c copy`. It shall not emit conversion-style per-stream `-map` or
+  `-c:*` options that could drop, reorder, or transcode streams during a
+  metadata-only edit.
+- `METADATA_EDITOR-0014`: Because ffmpeg cannot rewrite the source file in
+  place, `ffx edit` shall write to a temporary output file on the same
+  filesystem as the source file and shall replace the original path only after
+  ffmpeg reports success.
+- `METADATA_EDITOR-0015`: The temporary output path shall preserve the original
+  container type and file extension. The feature shall not silently change the
+  container or extension during a metadata-only edit.
+- `METADATA_EDITOR-0016`: If the rewrite step fails, the original file shall
+  remain untouched. The system shall not leave the user with a partially
+  replaced source file.
+- `METADATA_EDITOR-0017`: After a successful replace, the application shall
+  reprobe the rewritten file, refresh the baseline descriptor from disk, reset
+  the draft state to that fresh baseline, and clear the dirty state.
+- `METADATA_EDITOR-0018`: Edit mode shall track whether unsaved draft changes
+  exist and shall require confirmation before dismissing the screen or quitting
+  the app when such changes would be lost.
+- `METADATA_EDITOR-0019`: Edit mode shall not inject conversion-only encoding
+  metadata such as encoder quality or preset markers.
+- `METADATA_EDITOR-0020`: Signature-tag behavior shall be explicit for
+  metadata-only editing. The default behavior shall not add a misleading
+  recoding-style signature to a file that was only remuxed for metadata
+  updates.
+- `METADATA_EDITOR-0021`: Configured metadata-removal rules from the local
+  configuration shall be surfaced clearly in the UI and in the planned-changes
+  view. If those rules are applied during save, the operator shall be able to
+  tell that the file will be cleaned in addition to any manual edits.
+- `METADATA_EDITOR-0022`: Edit mode shall provide an in-screen operator toggle
+  for config-driven cleanup so a user can switch between pure manual metadata
+  edits and metadata edits plus configured tag cleanup without leaving the
+  editor.
+- `METADATA_EDITOR-0023`: The existing global `--dry-run` behavior shall apply
+  to `ffx edit`. In dry-run mode the command shall not replace the original
+  file and shall expose the planned write operation clearly enough for the user
+  to understand what would happen.
+- `METADATA_EDITOR-0024`: Every ffmpeg invocation performed by `ffx edit`
+  shall be surfaced to the operator as a notification in the edit UI.
+- `METADATA_EDITOR-0025`: When application verbosity is greater than zero, the
+  notification for an `ffx edit` ffmpeg invocation shall include the concrete
+  ffmpeg command line.
+
+## Acceptance
+
+- `ffx edit /path/to/file.mkv` opens successfully on a workstation where the
+  configured database is missing, empty, incompatible, or intentionally
+  inaccessible.
+- Opening a file in edit mode does not trigger database bootstrap or migration
+  prompts.
+- A user can change a container tag, save, and see the rewritten file at the
+  same path with the updated metadata.
+- A user can change a stream title or language, save, and see the rewritten
+  file at the same path with the updated stream metadata.
+- A user can change `default` or `forced` on a track, save, and see the
+  rewritten file at the same path with the updated dispositions.
+- The planned-changes view reflects manual edits relative to the original file
+  and, when enabled, any configured cleanup removals.
+- No rendered Rich or Textual color markup appears in the saved file metadata.
+- Saving metadata with files that contain PGS subtitle tracks or other
+  non-text subtitle codecs preserves those streams instead of dropping them.
+- If ffmpeg fails while saving, the original file remains present and readable
+  at the original path.
+- In dry-run mode, the original file remains untouched.
+
+## Current Code Fit
+
+- Good fit:
+  - `FfxController.runJob(...)` already has a `VideoEncoder.COPY` path that
+    can remux streams and apply metadata and disposition tokens.
+  - `MediaDescriptorChangeSet` already computes container-tag, stream-tag, and
+    disposition differences and can generate ffmpeg metadata tokens.
+  - `TagDetailsScreen` and `TrackDetailsScreen` already provide reusable edit
+    dialogs for draft state.
+  - `PatternDetailsScreen` already demonstrates add, edit, and delete flows for
+    tags and tracks in a draft-first UI.
+- Refactor required:
+  - `ffx` CLI initialization currently creates a database context for all
+    non-lightweight commands, so `edit` needs its own DB-free bootstrap path.
+  - `FileProperties` currently instantiates `PatternController` eagerly, so
+    probing must be split from pattern matching or made lazy.
+  - `MediaDetailsScreen` currently assumes `command == 'inspect'` and mixes
+    file state with database-backed target-pattern state.
+  - `MediaDetailsScreen` currently mutates the probed source descriptor
+    directly. Edit mode needs an immutable baseline descriptor and a separate
+    mutable draft descriptor.
+  - `TrackDetailsScreen` currently exposes structural fields that are valid for
+    pattern authoring but too dangerous for metadata-only file editing.
+
+## Risks
+
+- Container-level metadata support differs across formats, so some requested tag
+  changes may not round-trip identically through ffmpeg for every supported
+  container.
+- The existing metadata-removal implementation is conversion-oriented and may
+  remove tags more aggressively than a user expects from a manual editor unless
+  cleanup policy is made explicit.
+- The current codebase lacks a dedicated descriptor clone API, so draft-state
+  separation should be implemented deliberately instead of via accidental shared
+  references.
+- Replacing a file path with a temporary output changes inode identity, so any
+  future requirement around preserving timestamps, hard links, or extended
+  attributes would need additional explicit handling.

requirements/pattern_management.md

@@ -0,0 +1,68 @@
+# Pattern Management
+
+This file defines the behavioral contract for managing shows, patterns, and
+pattern-backed filename matching.
+
+Primary source: actual tool code in `src/ffx/`.
+Secondary source: operator intent captured in task discussion.
+
+## Scope
+
+- The show, pattern, and track hierarchy stored in SQLite.
+- The role of a pattern as a reusable normalization definition for related media files.
+- Filename-driven assignment of a scanned media file to one show through one matching pattern.
+- Duplicate-match handling when more than one pattern matches the same filename.
+
+## Terms
+
+- `show`: logical series identity such as one TV show entry in the database.
+- `pattern`: regex-backed normalization definition attached to one show.
+- `track`: one persisted target-track definition attached to one pattern.
+- `scanned media file`: one source file currently being inspected or converted.
+- `duplicate pattern match`: a filename state where more than one stored pattern matches the same scanned media file.
+- `pattern-backed target schema`: the combination of one pattern's stored media tags and stored track definitions.
+
+## Rules
+
+- `PATTERN_MANAGEMENT-0001`: The domain model shall treat a show as the parent entity for patterns that describe distinct release families or normalization schemas for that show. A show may temporarily exist without patterns during editing or initial TUI creation.
+- `PATTERN_MANAGEMENT-0002`: Each persisted pattern shall belong to exactly one show.
+- `PATTERN_MANAGEMENT-0003`: The domain model shall treat a pattern as the reusable normalization definition for a series of media files expected to share the same internal track layout and materially similar stream and container metadata.
+- `PATTERN_MANAGEMENT-0004`: Each persisted track definition shall belong to exactly one pattern.
+- `PATTERN_MANAGEMENT-0005`: A pattern may also carry pattern-level media tags. The pattern's media tags plus its track definitions together form the pattern-backed target schema.
+- `PATTERN_MANAGEMENT-0006`: A scanned media file shall resolve to at most one pattern and therefore at most one show.
+- `PATTERN_MANAGEMENT-0007`: If no pattern matches a filename, the file shall remain unmatched rather than being assigned implicitly.
+- `PATTERN_MANAGEMENT-0008`: If more than one pattern matches the same filename, the system shall raise a duplicate pattern match error instead of silently selecting one.
+- `PATTERN_MANAGEMENT-0009`: Duplicate-match detection shall apply regardless of whether the competing patterns belong to the same show or to different shows.
+- `PATTERN_MANAGEMENT-0010`: Exact duplicate pattern definitions for the same show should not create multiple persisted pattern rows.
+- `PATTERN_MANAGEMENT-0011`: A persisted pattern shall define one or more tracks. Creating or retaining a zero-track pattern in the database is invalid managed state and shall be prohibited.
+- `PATTERN_MANAGEMENT-0012`: A show may exist without patterns as an intermediate editing state, for example when a user creates the show first in the TUI and adds patterns later.
+- `PATTERN_MANAGEMENT-0013`: Operator-facing pattern management should expose the owning show, regex pattern, stored track set, and stored media-tag set so a user can reason about matching and normalization behavior.
+- `PATTERN_MANAGEMENT-0014`: Matching semantics shall be deterministic and documented. Implicit "last matching pattern wins" behavior is not acceptable released behavior.
+
+## Acceptance
+
+- A filename that matches exactly one pattern yields one matched pattern and one show identity.
+- A filename that matches no pattern yields no matched pattern and an unmatched state.
+- A filename that matches more than one pattern yields an explicit duplicate-match error.
+- A pattern-backed target schema can be reconstructed from one pattern's stored media tags and stored track definitions.
+- A show may be stored before any patterns are attached to it.
+- A pattern cannot be stored or retained as a valid managed pattern unless at least one track is defined for it.
+- Pattern-backed conversion never proceeds with two competing matching patterns for the same input filename.
+
+## Current Code Fit
+
+- `src/ffx/model/show.py` implements a one-to-many `Show -> Pattern` relationship.
+- `src/ffx/model/pattern.py` implements `Pattern.show_id`, a one-to-many `Pattern -> Track` relationship, a one-to-many `Pattern -> MediaTag` relationship, and a unique `(show_id, pattern)` constraint for freshly created databases.
+- `src/ffx/model/track.py` implements `Track.pattern_id`, so each persisted track belongs to one pattern.
+- `src/ffx/model/pattern.py` reconstructs a pattern-backed target schema through `Pattern.getMediaDescriptor(...)`, combining stored media tags and stored tracks.
+- `src/ffx/file_properties.py` assumes a scanned file resolves to at most one pattern, because it stores only one `self.__pattern` and derives one `show_id` from it.
+- `src/ffx/pattern_controller.py` prevents exact duplicate `(show_id, pattern)` definitions during create and update flows, and it refreshes cached compiled regexes when stored pattern expressions change.
+- `src/ffx/pattern_controller.py` now complies with duplicate-match safety. `matchFilename(...)` scans deterministically, returns exactly one match, returns `{}` for no match, and raises an explicit duplicate-pattern-match error when more than one pattern matches the same filename.
+- The current persistence layer already aligns with the intended empty-show workflow because a show can exist without patterns.
+- New pattern creation and schema replacement flows now require at least one track, and `TrackController.deleteTrack(...)` prevents deleting the last persisted track from a pattern.
+- Trackless legacy rows can still exist in preexisting databases, but matching now rejects them explicitly instead of letting them participate silently.
+
+## Risks
+
+- The intended "release family" meaning of a pattern is a domain assumption, not something the code verifies automatically across all files matching that pattern.
+- Preexisting databases created before the newer validation rules may still contain invalid rows, so upgrade and cleanup paths should continue to treat explicit validation failures as recoverable operator signals.

requirements/project.md

@@ -0,0 +1,124 @@
+## Purpose And Scope
+
+- Project name: FFX
+- User problem: TV episode files from mixed sources arrive with inconsistent codecs, stream metadata, subtitle layouts, season and episode numbering, and output filenames, which makes them awkward to archive and use in media-player applications.
+- Target users: Individual operators curating a local TV media library on a workstation, especially users willing to define normalization rules per show.
+- Success outcome: A user can inspect source files, define reusable show and pattern rules, and produce output files whose streams, metadata, and filenames follow a predictable schema for web playback and library import.
+- Out of scope:
+  - Multi-user or hosted service workflows.
+  - General movie-library management.
+  - Distributed transcoding or remote job orchestration.
+  - Broad media-server administration beyond file preparation.
+
+## Required Product
+
+- Deliverable type: Installable Python command-line application with a Textual terminal UI for inspection and rule editing.
+- Core capabilities:
+  - Maintain an SQLite-backed database of shows, filename-matching patterns, per-pattern stream layouts and metadata tags, and optional season-shift rules.
+  - Inspect existing media files through `ffprobe` and compare discovered stream metadata with stored normalization rules.
+  - Convert media files through `ffmpeg` into a normalized output layout, including video recoding, audio transcoding to Opus, metadata cleanup and rewrite, and controlled disposition flags.
+  - Build output filenames from detected or configured show, season, and episode information, optionally enriched from TMDB and a configurable Jinja-style filename template.
+  - Support auxiliary file operations such as subtitle import, unmuxing, crop detection, rename-only conversion runs, and direct in-place episode renaming.
+- Supported environments:
+  - Local execution on a Python-capable workstation.
+  - Best-supported on Linux-like systems because the implementation assumes `~/.local`, `/dev/null`, `nice`, and `cpulimit`.
+  - Requires `ffmpeg`, `ffprobe`, and `cpulimit` on `PATH`.
+- Operational owner: The local user running the tool and maintaining its config, database, and external tooling.
+
+## Suggested User Stories
+
+- As a library maintainer, I want to define show-specific matching rules once so that future source files can be normalized automatically.
+- As an operator, I want to inspect a file before conversion so that I can compare its actual streams and tags against the stored target schema.
+- As a user preparing web-playback files, I want to recode video and audio with a small set of predictable options so that results are compatible and consistently named.
+- As a user dealing with nonstandard releases, I want CLI overrides for language, title, stream order, default and forced tracks, and season and episode data so that one-off fixes do not require database edits first.
+- As a user importing anime or other shifted numbering schemes, I want season and episode offsets at the show level with optional pattern-specific overrides so that generated filenames align with TMDB and media-library expectations.
+
+## Functional Requirements
+
+- The system shall provide a CLI entrypoint named `ffx` with commands for `convert`, `inspect`, `shows`, `rename`, `unmux`, `cropdetect`, `setup`, `configure_workstation`, `upgrade`, `version`, and `help`.
+- The system shall support a two-step local installation and preparation flow:
+  - `tools/setup.sh` is the bootstrap entrypoint for the first step and shall own bundle virtualenv creation, package installation, shell alias exposure, and optional Python test-package installation.
+  - `tools/configure_workstation.sh` is the bootstrap entrypoint for the second step and shall own workstation dependency checks and installation plus local config and directory seeding.
+  - After the bundle is installed, `ffx setup` and `ffx configure_workstation` shall remain aligned wrapper entrypoints for those same two steps.
+- The CLI command `ffx setup` shall act as a wrapper for the first-step bundle-preparation flow in `tools/setup.sh`.
+- The CLI command `ffx configure_workstation` shall act as a wrapper for the second-step preparation flow in `tools/configure_workstation.sh`.
+- The system shall persist reusable normalization rules in SQLite for:
+  - shows and show formatting digits,
+  - optional show-level notes,
+  - optional show-level quality defaults,
+  - regex-based filename patterns,
+  - per-pattern media tags,
+  - per-pattern stream definitions,
+  - show-level and pattern-level shifted-season mappings,
+  - internal database version properties.
+- The system shall apply supported ordered database migrations automatically when opening an older local database file and shall fail fast when no supported path exists.
+- Before applying a required database migration, the system shall show the current version, target version, required sequential steps, and whether each corresponding migration module is present, then require user confirmation.
+- Before applying a confirmed file-backed database migration, the system shall create an in-place backup copy whose filename includes the covered version range.
+- Detailed show, pattern, and duplicate-match management rules live in `requirements/pattern_management.md`.
+- The system shall inspect source media using `ffprobe` and derive a structured description of container metadata and streams.
+- The system shall optionally open a Textual UI to browse shows, inspect files, and create, edit, or delete shows, patterns, stream definitions, tags, and shifted-season rules.
+- The system shall match filenames against stored regex patterns to decide whether an input file should inherit a target stream and metadata schema.
+- The system shall convert supported input files (`mkv`, `mp4`, `avi`, `flv`, `webm`) with `ffmpeg`, supporting at least:
+  - VP9, AV1, and H.264 video encoding,
+  - Opus audio encoding with bitrate selection based on channel layout,
+  - metadata and disposition rewriting,
+  - optional crop detection and crop application,
+  - optional deinterlacing and denoising,
+  - optional subtitle import from external files,
+  - rename-only move mode.
+- The system shall support optional TMDB lookups to resolve show names, years, and episode titles when a show ID, season, and episode are available.
+- The system shall generate output filenames from show metadata, season and episode indices, and episode names using the configured filename template.
+- The system shall allow CLI overrides for stream languages, stream titles, default and forced tracks, stream order, TMDB show and episode data, output directory, label prefix, and processing resource limits.
+- The system shall resolve encoding quality by precedence `CLI override -> pattern -> show -> encoder default` and shall report the chosen value and source.
+- The system shall resolve season shifting by precedence `pattern -> show -> identity default` and shall report the chosen mapping and source.
+- Processing resource limit rules:
+  - `--nice` shall accept niceness values from `-20` through `19`; omitting the option shall disable niceness adjustment.
+  - `--cpu` shall accept either a positive absolute `cpulimit` value such as `200`, or a percentage suffixed with `%` such as `25%` to represent a share of present CPUs; omitting the option or using `0` shall disable CPU limiting.
+  - When both limits are configured, the process wrapper shall execute the target command through `cpulimit` around a `nice -n ...` invocation so both limits apply to the launched media command.
+- The system shall support extracting streams into separate files via `unmux` and reporting suggested crop parameters via `cropdetect`.
+- The system shall support in-place episode renaming via `rename`, requiring a `--prefix`, accepting optional `--season` and `--suffix` overrides, preserving the source extension, and supporting dry-run output without moving files.
+- Crop detection shall use a configurable sampling window, defaulting to a 60-second seek and a 180-second analysis duration, and repeated crop-detection requests for the same source plus sampling window shall reuse cached results within one process.
+- The system shall handle invalid input and system failures gracefully by logging warnings or raising `click` errors for missing files, invalid media, missing TMDB credentials, incompatible database versions, and ambiguous track dispositions when prompting is disabled.
+
+## Quality Requirements
+
+- The system should stay understandable as a small local tool: controllers, descriptors, models, and screens should remain separate enough for contributors to trace a workflow end to end.
+- The system should produce predictable output for the same database rules, CLI overrides, and source files.
+- The system should preserve a lightweight operational footprint: local SQLite state, local log file, no mandatory background services.
+- The system should be testable through modern automatically discovered tests and through remaining legacy harness coverage during migration.
+- The system should expose enough logging to diagnose failed probes, failed conversions, and rule mismatches without requiring a debugger.
+
+## Constraints And Assumptions
+
+- Technology constraints:
+  - Python package built with setuptools.
+  - Primary libraries: `click`, `textual`, `sqlalchemy`, `jinja2`, `requests`.
+  - Conversion and inspection rely on external executables rather than pure-Python media libraries.
+- Hosting or infrastructure constraints:
+  - Intended for local execution, not server deployment.
+  - Stores default state in `~/.local/etc/ffx.json`, `~/.local/var/ffx/ffx.db`, and `~/.local/var/log/ffx.log`.
+- Timeline constraints:
+  - The current implemented scope reflects a compact alpha release stream up to version `0.3.1`.
+- Team capacity assumptions:
+  - Maintained as a small codebase where simple patterns and direct controller logic are preferred over framework-heavy abstractions.
+- Third-party dependencies:
+  - `ffmpeg`, `ffprobe`, and `cpulimit`.
+  - TMDB API access through `TMDB_API_KEY` for metadata enrichment.
+- Installation assumptions:
+  - The Python-side bundle install step and optional Python test extras are managed by `tools/setup.sh`, with `ffx setup` as the aligned wrapper after bootstrap.
+  - The workstation-preparation step is managed separately by `tools/configure_workstation.sh` or `ffx configure_workstation`.
+
+## Acceptance Scope
+
+- First release boundary:
+  - Local installation through `pip`.
+  - Working SQLite-backed rule storage.
+  - Functional CLI conversion and inspection workflows.
+  - Textual CRUD flows for shows, patterns, tags, tracks, and shifted seasons.
+  - TMDB-assisted filename generation, subtitle import, season shifting, database versioning, and configurable output filename templating.
+- Excluded follow-up ideas:
+  - Completing placeholder screens such as settings and help.
+  - Hardening platform portability beyond Linux-like systems.
+  - Broader media types, richer release packaging, and production-grade background processing.
+- Demonstration scenario:
+  - Inspect a TV episode file, define or update the matching show and pattern in the TUI, then run `ffx convert` so the result uses the stored stream schema, optional TMDB episode naming, and a normalized output filename.

requirements/shifted_seasons_handling.md

@@ -0,0 +1,177 @@
+# Numbering Mapping Handling
+
+This file defines the behavioral contract for mapping source season and episode
+numbering to target season and episode numbering through stored shifted-season
+rules.
+
+Primary sources:
+- `requirements/project.md`
+- `requirements/architecture.md`
+- actual tool code in `src/ffx/`
+
+Secondary source:
+- `SCRATCHPAD.md`, used only to clarify current hardening gaps and not as the
+  primary contract source.
+
+## Scope
+
+- Persisting shifted-season rules in SQLite.
+- Allowing shifted-season rules to be attached either to a show or to a
+  specific pattern.
+- Selecting at most one active shifted-season rule for one concrete source
+  season and episode tuple.
+- Applying additive season and episode offsets to produce target numbering.
+- Using shifted target numbering during `convert` for TMDB episode lookup and
+  generated season and episode filename tokens.
+- Managing show-level default mappings and pattern-level override mappings from
+  the Textual editing workflows.
+
+## Out Of Scope
+
+- General filename parsing rules for detecting season and episode values.
+- Standalone `rename` command behavior, which currently uses explicit rename
+  inputs rather than stored shifted-season rules.
+- Stream or track mapping behavior unrelated to season and episode numbering.
+
+## Terms
+
+- `shifted-season rule`: one persisted row describing how one source-numbering
+  range maps to target numbering through additive offsets.
+- `show-level shifted-season rule`: a rule attached directly to a show and used
+  as the fallback mapping layer for that show.
+- `pattern-level shifted-season rule`: a rule attached directly to a pattern and
+  used as the override mapping layer for that pattern.
+- `source numbering`: the season and episode values detected from the current
+  source file or supplied as source-side conversion inputs before shifting.
+- `target numbering`: the season and episode values after one active
+  shifted-season rule has been applied.
+- `original season`: the source-domain season number a shifted-season rule is
+  eligible to match.
+- `episode range`: the optional source-domain episode interval covered by one
+  shifted-season rule.
+- `open bound`: an unbounded start or end of the episode range. Current storage
+  uses `-1` as the internal sentinel for an open bound.
+- `active shifted-season rule`: the single rule selected for one concrete input
+  after precedence resolution.
+- `identity mapping`: the default `1:1` outcome where source numbering is used
+  unchanged.
+
+## Rules
+
+- `SHIFTED_SEASONS_HANDLING-0001`: The domain model shall allow a
+  shifted-season rule to be owned by exactly one of:
+  - one show
+  - one pattern
+- `SHIFTED_SEASONS_HANDLING-0002`: A single shifted-season rule shall not
+  belong to both a show and a pattern at the same time.
+- `SHIFTED_SEASONS_HANDLING-0003`: A shifted-season rule shall carry these
+  fields: `original_season`, `first_episode`, `last_episode`,
+  `season_offset`, and `episode_offset`.
+- `SHIFTED_SEASONS_HANDLING-0004`: `season_offset` and `episode_offset` shall
+  be additive signed integers applied to matched source numbering to produce
+  target numbering.
+- `SHIFTED_SEASONS_HANDLING-0005`: A shifted-season rule shall match a source
+  tuple only when:
+  - the source season equals `original_season`
+  - the source episode is greater than or equal to `first_episode` when the
+    lower bound is closed
+  - the source episode is less than or equal to `last_episode` when the upper
+    bound is closed
+- `SHIFTED_SEASONS_HANDLING-0006`: An open lower or upper episode bound shall
+  represent an unbounded side of the covered source episode range.
+- `SHIFTED_SEASONS_HANDLING-0007`: If one shifted-season rule matches, target
+  numbering shall be:
+  - `target season = source season + season_offset`
+  - `target episode = source episode + episode_offset`
+- `SHIFTED_SEASONS_HANDLING-0008`: If no shifted-season rule matches, source
+  numbering shall pass through unchanged.
+- `SHIFTED_SEASONS_HANDLING-0009`: Shifted-season handling shall operate in a
+  source-to-target numbering model. Stored rules map detected source numbering
+  to the target numbering used by conversion-facing metadata and output naming.
+- `SHIFTED_SEASONS_HANDLING-0010`: Pattern matching identifies the owning show
+  and optionally a more specific owning pattern. Resolution of the active
+  shifted-season rule shall use this precedence order:
+  - matching pattern-level rule
+  - matching show-level rule
+  - identity mapping
+- `SHIFTED_SEASONS_HANDLING-0011`: At most one shifted-season rule may be
+  active for one concrete source season and episode tuple. Shifted-season rules
+  shall never stack or compose.
+- `SHIFTED_SEASONS_HANDLING-0012`: Within one owner scope, shifted-season rules
+  shall not overlap in their effective episode coverage for the same
+  `original_season`.
+- `SHIFTED_SEASONS_HANDLING-0013`: If a shifted-season rule uses two closed
+  episode bounds, `last_episode` shall be greater than or equal to
+  `first_episode`.
+- `SHIFTED_SEASONS_HANDLING-0014`: Shifted-season rule evaluation shall be
+  deterministic. Released behavior shall not depend on arbitrary database row
+  order when invalid overlapping rules exist.
+- `SHIFTED_SEASONS_HANDLING-0015`: A pattern-level rule is permitted to map to
+  zero offsets. Such a rule is a valid explicit override that beats show-level
+  fallback and produces identity mapping for its covered source range.
+- `SHIFTED_SEASONS_HANDLING-0016`: During `convert`, when show, season, and
+  episode values are available and stored shifting is active, the shifted target
+  numbering shall drive:
+  - TMDB episode lookup
+  - season and episode filename tokens such as `S01E02`
+  - generated episode basenames that include season and episode numbering
+- `SHIFTED_SEASONS_HANDLING-0017`: When conversion is supplied explicit
+  target-domain season or episode values for TMDB naming, the system shall not
+  apply stored shifting on top of those already-targeted values.
+- `SHIFTED_SEASONS_HANDLING-0018`: Operator-facing editing shall expose
+  shifted-season rule management in both of these places:
+  - show editing for show-level default mappings
+  - pattern editing for pattern-level override mappings
+- `SHIFTED_SEASONS_HANDLING-0019`: User-facing shifted-season editing should
+  present open episode bounds as a natural empty-state input rather than forcing
+  operators to type the internal sentinel directly.
+
+## Acceptance
+
+- A show can exist with zero or more show-level shifted-season rules.
+- A pattern can exist with zero or more pattern-level shifted-season rules.
+- A shifted-season rule is stored against exactly one owner scope.
+- A source tuple matching a pattern-level rule yields target numbering from that
+  rule even when a matching show-level rule also exists.
+- A source tuple matching no pattern-level rule but matching a show-level rule
+  yields target numbering from the show-level rule.
+- A source tuple matching neither scope yields identity mapping.
+- A pattern-level zero-offset rule can explicitly override a nonzero show-level
+  rule for the same covered source range.
+- Two shifted-season rules for the same owner scope and original season cannot
+  both be valid if they cover overlapping episode ranges.
+- During `convert`, shifted numbering is what TMDB episode lookup and generated
+  season and episode tokens see when stored shifting is active.
+- The TUI can display and maintain shifted-season rules from both the show and
+  pattern editing flows.
+
+## Current Code Fit
+
+- `src/ffx/model/show.py` and `src/ffx/model/pattern.py` now both expose
+  shifted-season relationships, and `src/ffx/model/shifted_season.py` stores
+  each rule against exactly one owner scope through `show_id` or `pattern_id`.
+- `src/ffx/shifted_season_controller.py` now resolves mappings with
+  pattern-over-show precedence and applies at most one active rule for a source
+  tuple.
+- `src/ffx/show_details_screen.py`,
+  `src/ffx/shifted_season_details_screen.py`, and
+  `src/ffx/shifted_season_delete_screen.py` provide reusable shifted-season
+  editing dialogs, and `src/ffx/pattern_details_screen.py` now exposes the
+  pattern-level override flow.
+- `src/ffx/cli.py` now resolves shifted numbering during `convert` from:
+  pattern-level match, then show-level match, then identity mapping.
+- `src/ffx/database.py` now migrates version-2 databases to version 3 by
+  preserving existing show-level rows and extending the schema for pattern-level
+  ownership.
+
+## Risks
+
+- The current CLI groups `--show`, `--season`, and `--episode` under one
+  override bucket used for TMDB-related behavior. Source-domain versus
+  target-domain semantics of each override must stay documented clearly so
+  stored shifting is neither skipped nor double-applied unexpectedly.
+- Existing version-2 databases only contain show-owned shifted-season rows, so a
+  version-3 migration must preserve those rows as the show-level fallback layer.
+- Current modern automated test coverage for shifted-season behavior is light,
+  so precedence, migration, and convert-time numbering behavior need focused
+  tests.

requirements/source_file_formats.md

@@ -0,0 +1,90 @@
+# Source File Formats
+
+This file defines source-file-format-specific processing requirements for FFX.
+It is intended to grow as additional relevant source file types are identified.
+
+The first covered format is Matroska media that contains styled ASS/SSA
+subtitle streams together with embedded font attachments.
+
+## Scope
+
+- Detecting source files that use ASS subtitle streams together with embedded
+  font attachments needed for correct rendering.
+- Defining the required `ffx convert` behavior when this format is present.
+- Preserving the required attachment streams during conversion.
+- Keeping normal subtitle-track manipulation behavior for the ASS subtitle
+  tracks themselves.
+
+## Out Of Scope
+
+- General subtitle behavior for sources that do not carry this pattern.
+- A complete catalog of all source file formats FFX may support later.
+
+## Terms
+
+- `styled ASS source`: a source media file that contains one or more subtitle
+  streams with `codec_type="subtitle"` and `codec_name="ass"` together with
+  one or more font-bearing attachment streams.
+- `font attachment`: an attachment stream whose metadata identifies a font
+  payload, commonly through `tags.mimetype` and attachment filename metadata.
+- `external subtitle feed`: subtitle tracks supplied from separate subtitle
+  files through the existing subtitle-import path.
+- `special attachment subtracks`: the embedded font attachment streams that
+  belong to the styled ASS source pattern.
+
+## Rules
+
+- `SOURCE_FILE_FORMATS-0001`: The system shall recognize the styled ASS source
+  pattern.
+- `SOURCE_FILE_FORMATS-0002`: Recognition shall not depend on fixed stream
+  counts, fixed stream indices, or one exact attachment count.
+- `SOURCE_FILE_FORMATS-0003`: Recognition shall use the best available ffprobe
+  signals. For known subtitle streams this includes
+  `codec_type="subtitle"` together with `codec_name="ass"`.
+- `SOURCE_FILE_FORMATS-0004`: Recognition of the special attachment subtracks
+  shall use attachment-oriented signals such as `codec_type="attachment"` and
+  font-identifying metadata such as `tags.mimetype="font/ttf"` when present.
+- `SOURCE_FILE_FORMATS-0005`: Recognition shall tolerate known ffprobe
+  variation in attachment reporting, including files where attachment streams
+  do not expose a `codec_name` but do expose `codec_type="attachment"` and
+  font-identifying tags.
+- `SOURCE_FILE_FORMATS-0006`: When attachment metadata varies across files,
+  detection shall not depend on one exact MIME string alone. Detection shall
+  be written so the known pattern can vary while still recognizing font
+  attachments.
+- `SOURCE_FILE_FORMATS-0007`: When the styled ASS source pattern is detected,
+  `ffx convert` shall emit an operator-facing message that reports the
+  detection and hints that special subtitle preservation handling is being
+  applied.
+- `SOURCE_FILE_FORMATS-0008`: When the styled ASS source pattern is present on
+  the source file, `ffx convert` shall not process an external subtitle feed.
+  The command shall stop before conversion and report an error that explains
+  that separate subtitle-file import is incompatible with this source format.
+- `SOURCE_FILE_FORMATS-0009`: Normal manipulation of the ASS subtitle streams
+  themselves shall continue to work through the usual selection, ordering,
+  metadata, language, title, and disposition handling paths.
+- `SOURCE_FILE_FORMATS-0010`: The special attachment subtracks shall be
+  preserved in the target media file as-is rather than transcoded,
+  regenerated, or replaced from external sources.
+- `SOURCE_FILE_FORMATS-0011`: Preserving the special attachment subtracks
+  as-is includes retaining the attachment payload and the attachment metadata
+  required by consumers, especially attachment filename and mimetype
+  information.
+- `SOURCE_FILE_FORMATS-0012`: This file shall remain the extension point for
+  additional source-file-format contracts as FFX adds support for more special
+  source formats.
+
+## Acceptance
+
+- A source file matching the observed pattern of embedded ASS subtitles plus
+  font attachments is recognized even when the attachment streams do not carry
+  a `codec_name`.
+- `ffx convert` output contains a clear detection message before the actual
+  conversion work proceeds.
+- If external subtitle import is requested for such a source file, the command
+  fails fast with an explicit error instead of mixing sidecar subtitles into
+  the job.
+- Existing manipulation of the ASS subtitle tracks still works for metadata,
+  titles, languages, ordering, and dispositions.
+- The output media preserves the required font attachment streams and their
+  identifying metadata needed by downstream media players.

requirements/subtrack_mapping.md

@@ -0,0 +1,84 @@
+# Subtrack Mapping
+
+This file defines the behavioral contract for mapping input subtracks to output
+subtracks during conversion.
+
+Primary source: actual tool code in `src/ffx/`.
+Secondary source: `tests/legacy/`, used only to clarify intent and reveal gaps.
+
+## Scope
+
+- Ensuring each target subtrack is created from the corresponding source-subtrack information, including stream-level metadata.
+- Mapping input streams to output streams during conversion.
+- Using persisted pattern-track definitions from the database as the target schema.
+- Allowing omission and reordering of retained tracks.
+- Keeping stream-level metadata attached to the correct source-derived logical track after remapping.
+- Normalizing target output into ordered track groups: video, audio, subtitle, then special types such as fonts or images.
+
+## Terms
+
+- `source_index`: identity of the originating input stream from ffprobe or an imported source descriptor.
+- `index`: final output-track order across all retained tracks.
+- `sub_index`: per-type position within the retained tracks of one type, for example audio stream `0` or subtitle stream `1`.
+- `target schema`: stored or constructed output-track definition that decides which tracks are kept, omitted, reordered, and rewritten.
+- `separate source file`: additional file bound to one target track slot whose media payload replaces the regular source payload for that slot.
+
+## Rules
+
+- `SUBTRACK_MAPPING-0001`: The system shall represent source-stream identity separately from output order. `source_index`, `index`, and `sub_index` are distinct concepts and shall not be collapsed into one field.
+- `SUBTRACK_MAPPING-0002`: The system shall derive `source_index` for probed tracks from the original ffprobe stream index and preserve that identity through conversion planning.
+- `SUBTRACK_MAPPING-0003`: Pattern-backed track definitions stored in the database shall persist both target output order and originating source-stream identity.
+- `SUBTRACK_MAPPING-0004`: When a filename matches a pattern, the pattern target schema shall be the source of truth for which source tracks are retained, which are omitted, and in what order retained tracks appear in the output.
+- `SUBTRACK_MAPPING-0005`: A target track may refer only to an existing source track of the same type. Conversion shall fail fast when a target track refers to a nonexistent source stream or a source stream of a different type.
+- `SUBTRACK_MAPPING-0006`: The ffmpeg mapping phase shall be generated from target output order while resolving each retained output track back to its originating source stream via `source_index`.
+- `SUBTRACK_MAPPING-0007`: Reordering and omission shall preserve logical track identity. Stream-level metadata, titles, languages, and disposition decisions shall stay attached to the correct source-derived logical track after mapping.
+- `SUBTRACK_MAPPING-0008`: The system shall support one-off CLI stream-order overrides without requiring prior database edits.
+- `SUBTRACK_MAPPING-0009`: Operator-facing inspection and editing surfaces shall expose enough source-versus-target information to let a user reason about subtrack mapping decisions.
+- `SUBTRACK_MAPPING-0010`: Test coverage for subtrack mapping shall assert source-derived identity, omission, and output order explicitly. Final track counts or final type sequences alone are insufficient proof of correct mapping.
+- `SUBTRACK_MAPPING-0011`: Retained target tracks shall appear in ordered groups: video track or tracks first, then audio tracks, then subtitle tracks, then special types such as fonts or images. Within each group, the target schema shall define the order.
+- `SUBTRACK_MAPPING-0012`: Track omission is valid when required by output compatibility, when needed to normalize source tracks into the required target group order and schema, or when explicitly requested by database rules or CLI options.
+- `SUBTRACK_MAPPING-0013`: If source tracks do not already comply with the required target group order, conversion shall reorder retained tracks to match the target ordering contract without losing source-track identity or stream-level metadata lineage.
+
+## Separate Additional Source Files
+
+- `SUBTRACK_MAPPING-0014`: A separate source file may substitute the media payload of one target subtrack without changing that target track's intended output position.
+- `SUBTRACK_MAPPING-0015`: When a separate source file is used, the target track shall remain bound to the corresponding logical source track for mapping, validation, and metadata lineage.
+- `SUBTRACK_MAPPING-0016`: Metadata for a substituted target track shall be merged from the regular source track and the separate source file when available.
+- `SUBTRACK_MAPPING-0017`: If the separate source file provides a metadata field that is also present on the regular source track, the separate source file value shall win in the target output.
+- `SUBTRACK_MAPPING-0018`: If a metadata field is absent from the separate source file, the system shall fall back to the corresponding metadata from the regular source track or target schema rewrite rules.
+- `SUBTRACK_MAPPING-0019`: When `ffx convert` receives an explicit subtitle directory without a subtitle prefix, it shall discover sidecar files using the source media basename as the filename prefix.
+- `SUBTRACK_MAPPING-0020`: Basename-driven subtitle discovery shall first filter regular files by the exact `<source-basename>_` filename prefix and the configured subtitle extension.
+- `SUBTRACK_MAPPING-0021`: `--subtitle-extension` shall accept an extension with or without a leading dot, default to `vtt`, and apply to both basename-driven and explicit-prefix subtitle discovery.
+- `SUBTRACK_MAPPING-0022`: Basename-driven sidecar filenames shall identify the target subtitle track using the existing `<prefix>_<stream-index>_<language>[_<disposition>].<extension>` filename contract.
+- `SUBTRACK_MAPPING-0023`: A complete, valid basename-driven sidecar set shall proceed without confirmation and shall report the discovered substitutions to the operator.
+- `SUBTRACK_MAPPING-0024`: An incomplete but otherwise valid basename-driven sidecar set shall require confirmation before substituting only the represented subtitle tracks. `--yes` shall supply that confirmation without prompting. With `--no-prompt` and without `--yes`, conversion shall fail with an explanation instead.
+- `SUBTRACK_MAPPING-0025`: Basename-driven discovery shall fail before conversion when the filtered set contains too many files, malformed filenames, duplicate stream indices, or stream indices that do not identify subtitle tracks in the active media descriptor.
+
+## Acceptance
+
+- Given a source media descriptor and a pattern-backed target schema, the planned output tracks can be listed in final output order and each retained track can still be traced to one originating source stream.
+- Planned output order follows grouped target order: video, audio, subtitle, then special types.
+- Tracks not referenced by the target schema are omitted from output mapping.
+- Tracks may also be omitted when they are incompatible with the chosen output format or explicitly excluded by database or CLI rules.
+- Two retained target tracks never originate from the same source stream unless duplication is implemented explicitly as a separate feature.
+- If target-track metadata is rewritten after reordering, it is written onto the correct source-derived logical track rather than the track that merely occupies the same final output position.
+- Invalid target-to-source references fail deterministically before the conversion job is launched.
+- If a separate source file substitutes one target track, that track keeps its target slot and ordering while metadata is merged with separate-file values taking precedence when both sides provide the same field.
+- Given `A2_t01.mkv` and an explicit subtitle directory containing `A2_t01_2_deu_DEF.vtt`, `A2_t01_3_eng.vtt`, and `A2_t01_4_eng.vtt`, directory-only subtitle import recognizes and substitutes all three tracks without prompting.
+- Selecting `--subtitle-extension mkv` or `--subtitle-extension .mkv` selects the equivalent basename-matched `.mkv` sidecar set instead of the default `.vtt` set.
+- Given an incomplete but valid basename-matched sidecar set, `--yes` proceeds with only the represented subtitle substitutions, including when `--no-prompt` is also set.
+- A test proving subtrack mapping must assert at least one of: exact `source_index` to output-order mapping, omission of named source tracks, or preservation of per-track metadata after reorder.
+
+## Test Notes
+
+- `tests/legacy/scenario.py` names pattern behavior as `Filter/Reorder Tracks`.
+- `tests/legacy/scenario_4.py` is the strongest end-to-end signal because it runs DB-backed conversion and reapplies source indices before assertion.
+- `tests/legacy/track_tag_combinator_2_0.py` and `tests/legacy/track_tag_combinator_3_4.py` sort result tracks by `source_index` before checking tags, which matches the intended identity model.
+- Legacy permutation combinators define permutations but their assertion functions are stubs.
+- Some legacy scenarios produce `AP` and `SP` selectors but do not execute them.
+
+## Risks
+
+- `src/ffx/media_descriptor.py` contains an explicit `rearrangeTrackDescriptors()` path whose current implementation appears defective and under-tested.
+- Separate-source-file metadata precedence is only partly expressed in current implementation paths and should be covered directly in the rewritten test suite.
+- Production code expresses the mapping contract more clearly than the legacy harness, so a rewrite should add direct logic-level tests for mapping and reorder planning.

requirements/tests.md

@@ -0,0 +1,144 @@
+# Test Rewrite
+
+This file captures the structure executed by `tests/legacy_runner.py` today and
+defines the target shape for a complete rewrite.
+
+Detailed product rules for source-to-target subtrack mapping live in
+`requirements/subtrack_mapping.md`. This file describes only how tests cover
+that area.
+
+## Interpreter Requirement
+
+- Agents shall run Python-side test commands with `~/.local/share/ffx.venv/bin/python`.
+- This applies to the legacy harness, `unittest`, `pytest`, helper scripts, and `python -m ffx ...` test invocations.
+- Agents shall not silently substitute `python`, `python3`, or another interpreter for Python-side test work.
+- If `~/.local/share/ffx.venv/bin/python` is missing or not executable, agents shall stop and report the missing venv instead of continuing with Python-side test execution.
+
+## Shell Environment Requirement
+
+- Agents shall source `~/.bashrc` from an interactive Bash shell before running TMDB-dependent test commands or TMDB-dependent `python -m ffx ...` test invocations.
+- Agents shall not source `~/.bashrc.d/interactive/77_tmdb.sh` directly for normal test work; `~/.bashrc` is the required entry point.
+- In automation this means agents shall use an interactive Bash invocation such as `bash -ic 'source ~/.bashrc && ...'`, because a non-interactive `bash -lc` returns from `~/.bashrc` before the interactive fragments are loaded.
+- If sourcing `~/.bashrc` still does not provide required shell environment such as `TMDB_API_KEY`, agents shall stop and report the missing environment instead of continuing with TMDB-dependent test execution.
+
+## Current Harness
+
+- Entrypoint: `~/.local/share/ffx.venv/bin/python tests/legacy_runner.py run`
+- Runner style: custom Click CLI, not `pytest` or `unittest`
+- Commands:
+  - `run`: discover scenario files, instantiate each scenario, run yielded jobs
+  - `dupe`: helper command that creates duplicate media fixtures; not part of the test run
+- Filters: `--scenario`, `--variant`, `--limit`
+- Shared context:
+  - builds one mutable dict for the whole run
+  - installs loggers and writes `ffx_test_report.log`
+  - creates `ConfigurationController` eagerly
+  - tracks only passed and failed counters
+- Discovery:
+  - scenario files: `tests/legacy/scenario_*.py`
+  - combinators: `glob + importlib + inspect` by filename convention
+  - ordering: implicit glob order, no explicit sorting
+- Skip behavior:
+  - Scenario 4 is skipped when `TMDB_API_KEY` is missing
+  - only `TMDB_API_KEY_NOT_PRESENT_EXCEPTION` is caught at scenario construction time
+
+## Current Scenarios
+
+- `1`: `tests/legacy/scenario_1.py`
+  - focus: basename generation without pattern lookup or TMDB
+  - inputs per job: `1`
+  - jobs: `140`
+  - expected failures: `0`
+  - execution: build one synthetic source file, run `~/.local/share/ffx.venv/bin/python -m ffx convert`, assert filename selectors only
+  - selectors executed: `B`, `L`, `I`
+  - selectors defined but not executed: `S`, `R`
+- `2`: `tests/legacy/scenario_2.py`
+  - focus: conversion matrix over media layouts, dispositions, tags, and permutations
+  - inputs per job: `1`
+  - jobs: `8193`
+  - expected failures: `3267`
+  - execution: build one synthetic source file, run `~/.local/share/ffx.venv/bin/python -m ffx convert`, probe result with `FileProperties`, assert track layout and selected audio and subtitle metadata
+  - selectors executed: `M`, `AD`, `AT`, `SD`, `ST`
+  - selectors defined but not executed: `MT`, `AP`, `SP`, `J`
+- `4`: `tests/legacy/scenario_4.py`
+  - focus: pattern-driven batch conversion with SQLite state and live TMDB naming
+  - inputs per job: `6`
+  - jobs: `768`
+  - expected failures: `336`
+  - execution: build six synthetic preset files, recreate temp SQLite DB, insert show and pattern, run one batch convert command via `~/.local/share/ffx.venv/bin/python`, query TMDB during assertions
+  - selectors executed: `M`, `AD`, `AT`, `SD`, `ST`
+  - selectors defined but not executed: `MT`, `AP`, `SP`, `J`
+  - notes:
+    - uses `MediaCombinator6` only
+    - issues live HTTP requests through `TmdbController` with no request cache
+
+## Current Combinator Families
+
+- scenario files discovered: `3`
+- basename combinators discovered: `2`
+- media combinators discovered: `8`
+- media tag combinators discovered: `3`
+- disposition combinator 2 variants: `4`
+- disposition combinator 3 variants: `5`
+- track tag combinator 2 variants: `4`
+- track tag combinator 3 variants: `5`
+- indicator variants: `7`
+- label variants: `2`
+- show variants: `3`
+- release variants: `3`
+- permutation 2 variants: `2`
+- permutation 3 variants: `3`
+
+## Current Totals
+
+- full run without TMDB: `8333`
+- full run with TMDB: `9101`
+- Scenario 4 generated source files: `4608`
+- Scenario 4 live TMDB episode queries: `4608`
+
+## Current Behavior Areas
+
+- output basename rules for label, season and episode indicator, show name, and release suffix combinations
+- track layout normalization across the eight media combinator shapes from `VA` through `VAASSS`
+- two-track and three-track disposition edge cases, including intentional failure cases
+- two-track and three-track track-tag preservation checks, including checks that sort results by source identity
+- container-level media tag handling
+- pattern-backed conversion against a temporary SQLite database
+- TMDB-assisted episode naming for batch conversion
+
+## Structural Findings
+
+- The suite is process-heavy: most jobs run `ffmpeg` to generate a fixture and then spawn the FFX CLI as a subprocess.
+- The suite is integration-first and has almost no isolated unit-level coverage for pure logic.
+- The base `Combinator` class is a placeholder and is not the real abstraction boundary used by the suite.
+- Many combinator methods are placeholders: there are `25` `pass` statements across the current test modules.
+- Several assertion families are never executed because scenario selector dispatch is incomplete.
+- Scenario comments mention a Scenario 3, but no `scenario_3.py` exists.
+- `tests/legacy/_basename_combinator_1.py` is effectively orphaned because discovery only matches `basename_combinator_*.py`.
+- `tests/legacy/disposition_combinator_2_3 .py` contains an embedded space in the filename and is still part of discovery.
+- Expected failures are validated only as subprocess return-code matches, not as specific error types or messages.
+- The current suite depends on `ffmpeg`, `ffprobe`, SQLite, the local Python environment, and for Scenario 4 a live TMDB API key plus network access.
+
+## Rewrite Target
+
+- Replace the custom Click harness with a standard test runner, preferably `pytest`.
+- Split the suite into explicit layers: unit, integration, and optional external-system tests.
+- Keep unit tests as the default path and make them runnable without `ffmpeg`, `ffprobe`, TMDB, or a user config directory.
+- Model discovery explicitly in code instead of relying on glob-plus-reflection naming conventions.
+- Convert the current Cartesian-product combinators into readable parametrized cases grouped by behavior area.
+- Preserve the current behavior areas, but represent them with targeted cases instead of thousands of opaque variant IDs.
+- Make every assertion family explicit and executable; there must be no selector that is produced but never consumed.
+- Replace live TMDB access with fixtures or mocks in normal runs; any live-contract test must be opt-in.
+- Replace ad hoc subprocess return-code checks with assertions on typed exceptions, stderr content, or structured outputs.
+- Provide small reusable media fixtures or fixture builders so only a narrow integration slice needs `ffmpeg`-generated media.
+- Make database tests self-contained and fast through temporary databases and direct controller-level assertions.
+- Make ordering, naming, and selection deterministic so a contributor can predict exactly what will run.
+- Expose a small smoke suite for quick local runs and CI, plus a separately marked slower integration suite.
+- Prefer domain-oriented test modules over combinator-family modules: basename, pattern matching, metadata rewrite, track ordering, TMDB naming, CLI smoke, and failure handling.
+
+## Rewrite Acceptance
+
+- A default local test run finishes quickly and without network access.
+- A contributor can identify which behavior a failing test covers without decoding variant strings like `VAASSS-A:D10-S:T001`.
+- All current intended failure behaviors remain covered, but each one is asserted directly and readably.
+- The rewritten suite can be adopted by CI without requiring live TMDB credentials.

src/ffx/cli.py

@@ -41,13 +41,17 @@ CPU_OPTION_HELP = (
     + "Omit to disable; 0 also disables."
 )
 SUBTITLE_DIRECTORY_OPTION_HELP = (
-    "Load subtitles from here. When omitted and --subtitle-prefix is set, "
+    "Load subtitles from here. Without --subtitle-prefix, match the source filename "
+    + "basename. When omitted and --subtitle-prefix is set, "
     + "FFX uses the configured subtitlesDirectory base path plus the prefix as a subdirectory."
 )
 SUBTITLE_PREFIX_OPTION_HELP = (
     "Subtitle filename prefix. Requires --subtitle-directory, or a configured "
     + "subtitlesDirectory base path that contains a matching <prefix>/ subdirectory."
 )
+SUBTITLE_EXTENSION_OPTION_HELP = (
+    "External subtitle filename extension. A leading dot is optional."
+)
 UNMUX_OUTPUT_DIRECTORY_OPTION_HELP = (
     "Write extracted streams here. When omitted together with --subtitles-only and "
     + "--label, FFX uses the configured subtitlesDirectory base path plus the label."
@@ -96,6 +100,18 @@ def normalizeCpuOption(ctx, param, value):
         raise click.BadParameter(str(ex)) from ex
 
 
+def normalizeSubtitleExtension(ctx, param, value):
+    normalizedExtension = str(value).strip().lower()
+    if normalizedExtension.startswith('.'):
+        normalizedExtension = normalizedExtension[1:]
+    if not normalizedExtension or not normalizedExtension.isalnum():
+        raise click.BadParameter(
+            "Subtitle extension must contain only letters and numbers, "
+            + "with an optional leading dot."
+        )
+    return normalizedExtension
+
+
 def parseCutOptionValue(value) -> tuple[int, int] | None:
     if value is None:
         return None
@@ -146,11 +162,21 @@ def resolveSubtitleImportOptions(context, subtitleDirectory, subtitlePrefix):
         else ''
     )
 
-    if not resolvedSubtitlePrefix:
-        return False, resolvedSubtitleDirectory, resolvedSubtitlePrefix
-
     if resolvedSubtitleDirectory:
-        return True, resolvedSubtitleDirectory, resolvedSubtitlePrefix
+        if not os.path.isdir(resolvedSubtitleDirectory):
+            raise click.ClickException(
+                "The provided subtitle directory does not exist: "
+                + resolvedSubtitleDirectory
+            )
+        return (
+            True,
+            resolvedSubtitleDirectory,
+            resolvedSubtitlePrefix,
+            not resolvedSubtitlePrefix,
+        )
+
+    if not resolvedSubtitlePrefix:
+        return False, resolvedSubtitleDirectory, resolvedSubtitlePrefix, False
 
     configuredSubtitlesBaseDirectory = context['config'].getSubtitlesDirectoryPath()
     if not configuredSubtitlesBaseDirectory:
@@ -170,7 +196,85 @@ def resolveSubtitleImportOptions(context, subtitleDirectory, subtitlePrefix):
             + resolvedSubtitleDirectory
         )
 
-    return True, resolvedSubtitleDirectory, resolvedSubtitlePrefix
+    return True, resolvedSubtitleDirectory, resolvedSubtitlePrefix, False
+
+
+def importExternalSubtitles(
+    context,
+    mediaDescriptor,
+    sourceFileBasename,
+    season,
+    episode,
+    preserveDispositions=False,
+):
+    matchSourceBasename = context['subtitle_match_source_basename']
+    subtitlePrefix = (
+        sourceFileBasename
+        if matchSourceBasename
+        else context['subtitle_prefix']
+    )
+
+    try:
+        importResult = mediaDescriptor.importSubtitles(
+            context['subtitle_directory'],
+            subtitlePrefix,
+            season,
+            episode,
+            preserve_dispositions=preserveDispositions,
+            extension=context['subtitle_extension'],
+            strict=matchSourceBasename,
+        )
+    except (OSError, ValueError) as ex:
+        raise click.ClickException(
+            f"External subtitle discovery failed for '{sourceFileBasename}': {ex}"
+        ) from ex
+
+    if not matchSourceBasename:
+        return importResult
+
+    importedTrackIndices = importResult['imported_track_indices']
+    missingTrackIndices = importResult['missing_track_indices']
+    extension = context['subtitle_extension']
+    importedDescription = (
+        ', '.join(f"#{index}" for index in importedTrackIndices)
+        if importedTrackIndices
+        else 'none'
+    )
+    click.echo(
+        f"External subtitle scan for '{sourceFileBasename}': found "
+        + f"{importResult['candidate_count']} .{extension} file(s); "
+        + f"matched subtitle tracks {importedDescription}."
+    )
+
+    if not missingTrackIndices:
+        return importResult
+
+    missingDescription = ', '.join(f"#{index}" for index in missingTrackIndices)
+    incompleteMessage = (
+        f"External subtitle files are missing for subtitle tracks "
+        + f"{missingDescription} in '{sourceFileBasename}'."
+    )
+    if context.get('yes', False):
+        click.echo(
+            incompleteMessage
+            + " Continuing with the matching subtitle files because --yes is set."
+        )
+        return importResult
+
+    if context['no_prompt']:
+        raise click.ClickException(
+            incompleteMessage
+            + " Partial subtitle substitution requires confirmation, but --no-prompt is set."
+        )
+
+    click.echo(incompleteMessage)
+    if not click.confirm(
+        "Continue and substitute only the subtitle tracks with matching files?",
+        default=False,
+    ):
+        raise click.ClickException("External subtitle substitution aborted by user.")
+
+    return importResult
 
 
 def resolveUnmuxOutputDirectory(context, outputDirectory, subtitlesOnly, label):
@@ -181,7 +285,10 @@ def resolveUnmuxOutputDirectory(context, outputDirectory, subtitlesOnly, label):
     )
     resolvedLabel = str(label).strip()
 
-    if resolvedOutputDirectory or not subtitlesOnly or not resolvedLabel:
+    if resolvedOutputDirectory:
+        return resolvedOutputDirectory, True
+
+    if not subtitlesOnly or not resolvedLabel:
         return resolvedOutputDirectory, False
 
     configuredSubtitlesBaseDirectory = context['config'].getSubtitlesDirectoryPath()
@@ -194,6 +301,59 @@ def resolveUnmuxOutputDirectory(context, outputDirectory, subtitlesOnly, label):
     return os.path.join(configuredSubtitlesBaseDirectory, resolvedLabel), True
 
 
+def confirmUnmuxOutputDirectoryCreation(outputDirectory):
+    message = (
+        "Create unmux output directory and missing parents: "
+        + str(outputDirectory)
+    )
+
+    if not sys.stdin.isatty():
+        return click.confirm(message, default=True)
+
+    click.echo(f"{message} [Y/n]: ", nl=False)
+    while True:
+        char = click.getchar()
+        if char in ('\r', '\n'):
+            click.echo()
+            return True
+
+        normalizedChar = char.lower()
+        if normalizedChar == 'y':
+            click.echo(char)
+            return True
+        if normalizedChar == 'n':
+            click.echo(char)
+            return False
+        if char in ('\x03', '\x04'):
+            raise click.Abort()
+
+        click.echo("\nPlease respond with 'y' or 'n': ", nl=False)
+
+
+def ensureUnmuxOutputDirectory(context, outputDirectory):
+    resolvedOutputDirectory = os.path.expanduser(str(outputDirectory).strip())
+    if not resolvedOutputDirectory:
+        return False
+
+    if os.path.isdir(resolvedOutputDirectory):
+        return False
+
+    if os.path.exists(resolvedOutputDirectory):
+        raise click.ClickException(
+            "Unmux output path exists but is not a directory: "
+            + resolvedOutputDirectory
+        )
+
+    if context.get('dry_run', False):
+        return False
+
+    if not confirmUnmuxOutputDirectoryCreation(resolvedOutputDirectory):
+        raise click.ClickException("Unmux output directory creation aborted by user.")
+
+    os.makedirs(resolvedOutputDirectory, exist_ok=True)
+    return True
+
+
 def resolveIndicatorDigitLengths(context=None, showDescriptor=None):
     from ffx.show_descriptor import ShowDescriptor
 
@@ -753,14 +913,14 @@ def unmux(ctx,
     ctx.obj['resource_limits']['cpu_limit'] = cpu
     ctx.obj['resource_limits']['cpu_percent'] = cpu
 
-    output_directory, create_output_directory = resolveUnmuxOutputDirectory(
+    output_directory, requires_output_directory = resolveUnmuxOutputDirectory(
         ctx.obj,
         output_directory,
         subtitles_only,
         label,
     )
-    if create_output_directory and existingSourcePaths and not ctx.obj.get('dry_run', False):
+    if requires_output_directory and existingSourcePaths:
-        os.makedirs(output_directory, exist_ok=True)
+        ensureUnmuxOutputDirectory(ctx.obj, output_directory)
 
     shiftedSeasonController = ShiftedSeasonController(ctx.obj)
 
@@ -974,6 +1134,14 @@ def checkUniqueDispositions(context, mediaDescriptor: MediaDescriptor):
 
 @click.option('--subtitle-directory', type=str, default='', help=SUBTITLE_DIRECTORY_OPTION_HELP)
 @click.option('--subtitle-prefix', type=str, default='', help=SUBTITLE_PREFIX_OPTION_HELP)
+@click.option(
+    '--subtitle-extension',
+    type=str,
+    default='vtt',
+    callback=normalizeSubtitleExtension,
+    show_default=True,
+    help=SUBTITLE_EXTENSION_OPTION_HELP,
+)
 
 @click.option('--language', type=str, multiple=True, help='Set stream language. Use format <stream index>:<3 letter iso code>')
 @click.option('--title', type=str, multiple=True, help='Set stream title. Use format <stream index>:<title>')
@@ -1034,6 +1202,12 @@ def checkUniqueDispositions(context, mediaDescriptor: MediaDescriptor):
 @click.option("--dont-pass-dispositions", is_flag=True, default=False)
 
 @click.option("--no-prompt", is_flag=True, default=False)
+@click.option(
+    "--yes",
+    is_flag=True,
+    default=False,
+    help="Confirm partial external subtitle substitution without prompting.",
+)
 @click.option("--no-signature", is_flag=True, default=False)
 @click.option("--keep-mkvmerge-metadata", is_flag=True, default=False)
 
@@ -1070,6 +1244,7 @@ def convert(ctx,
 
             subtitle_directory,
             subtitle_prefix,
+            subtitle_extension,
 
             language,
             title,
@@ -1108,6 +1283,7 @@ def convert(ctx,
             no_pattern,
             dont_pass_dispositions,
             no_prompt,
+            yes,
             no_signature,
             keep_mkvmerge_metadata,
             
@@ -1162,6 +1338,7 @@ def convert(ctx,
     context['use_tmdb'] = not no_tmdb
     context['use_pattern'] = not no_pattern
     context['no_prompt'] = no_prompt
+    context['yes'] = yes
     context['no_signature'] = no_signature
     context['keep_mkvmerge_metadata'] = keep_mkvmerge_metadata
 
@@ -1180,6 +1357,7 @@ def convert(ctx,
         context['import_subtitles'],
         resolvedSubtitleDirectory,
         resolvedSubtitlePrefix,
+        context['subtitle_match_source_basename'],
     ) = resolveSubtitleImportOptions(
         context,
         subtitle_directory,
@@ -1188,6 +1366,7 @@ def convert(ctx,
     if context['import_subtitles']:
         context['subtitle_directory'] = resolvedSubtitleDirectory
         context['subtitle_prefix'] = resolvedSubtitlePrefix
+        context['subtitle_extension'] = subtitle_extension
 
 
     existingSourcePaths = [p for p in paths if os.path.isfile(p) and p.split('.')[-1] in SUPPORTED_INPUT_FILE_EXTENSIONS]
@@ -1431,10 +1610,13 @@ def convert(ctx,
             currentShowDescriptor = None
 
             if context['import_subtitles']:
-                sourceMediaDescriptor.importSubtitles(context['subtitle_directory'],
+                importExternalSubtitles(
-                                                        context['subtitle_prefix'],
+                    context,
-                                                        showSeason,
+                    sourceMediaDescriptor,
-                                                        showEpisode)
+                    sourceFileBasename,
+                    showSeason,
+                    showEpisode,
+                )
 
             if cliOverrides:
                 sourceMediaDescriptor.applyOverrides(cliOverrides)
@@ -1478,11 +1660,14 @@ def convert(ctx,
 
 
             if context['import_subtitles']:
-                targetMediaDescriptor.importSubtitles(context['subtitle_directory'],
+                importExternalSubtitles(
-                                                        context['subtitle_prefix'],
+                    context,
-                                                        showSeason,
+                    targetMediaDescriptor,
-                                                        showEpisode,
+                    sourceFileBasename,
-                                                        preserve_dispositions=True)
+                    showSeason,
+                    showEpisode,
+                    preserveDispositions=True,
+                )
 
             # ctx.obj['logger'].debug(f"tmd subindices: {[t.getIndex() for t in targetMediaDescriptor.getAllTrackDescriptors()]} {[t.getSubIndex() for t in targetMediaDescriptor.getAllTrackDescriptors()]}  {[t.getDispositionFlag(TrackDisposition.DEFAULT) for t in targetMediaDescriptor.getAllTrackDescriptors()]}")
             ctx.obj['logger'].debug(f"tmd subindices: {[t.getIndex() for t in targetMediaDescriptor.getTrackDescriptors()]} {[t.getSubIndex() for t in targetMediaDescriptor.getTrackDescriptors()]}  {[t.getDispositionFlag(TrackDisposition.DEFAULT) for t in targetMediaDescriptor.getTrackDescriptors()]}")

src/ffx/media_descriptor.py

@@ -431,10 +431,13 @@ class MediaDescriptor:
             importedFilePath = td.getExternalSourceFilePath()
 
             if importedFilePath:
-
+                substitutionMessage = (
-                self.__logger.info(f"Substituting subtitle stream #{td.getIndex()} "
+                    f"Substituting subtitle stream #{td.getIndex()} "
-                                    + f"({td.getType().label()}:{td.getSubIndex()}) "
+                    + f"({td.getType().label()}:{td.getSubIndex()}) "
-                                    + f"with import from file {td.getExternalSourceFilePath()}")
+                    + f"with import from file {td.getExternalSourceFilePath()}"
+                )
+                click.echo(substitutionMessage)
+                self.__logger.debug(substitutionMessage)
 
                 importFileTokens += [
                     "-i",
@@ -524,66 +527,153 @@ class MediaDescriptor:
         return inputMappingTokens
 
 
-    def searchSubtitleFiles(self, searchDirectory, prefix):
+    def searchSubtitleFiles(
+        self,
+        searchDirectory,
+        prefix,
+        extension=SUBTITLE_FILE_EXTENSION,
+        strict=False,
+    ):
 
-        sesld_match = re.compile(f"{prefix}_{MediaDescriptor.SEASON_EPISODE_STREAM_LANGUAGE_DISPOSITIONS_MATCH}")
+        normalizedExtension = str(extension).strip().lower()
-        sld_match = re.compile(f"{prefix}_{MediaDescriptor.STREAM_LANGUAGE_DISPOSITIONS_MATCH}")
+        if normalizedExtension.startswith('.'):
+            normalizedExtension = normalizedExtension[1:]
+        escapedPrefix = re.escape(prefix)
+        sesld_match = re.compile(
+            f"{escapedPrefix}_{MediaDescriptor.SEASON_EPISODE_STREAM_LANGUAGE_DISPOSITIONS_MATCH}"
+        )
+        sld_match = re.compile(
+            f"{escapedPrefix}_{MediaDescriptor.STREAM_LANGUAGE_DISPOSITIONS_MATCH}"
+        )
 
         subtitleFileDescriptors = []
+        subtitleFilenames = []
 
-        for subtitleFilename in os.listdir(searchDirectory):
+        for subtitleFilename in sorted(os.listdir(searchDirectory)):
-            if subtitleFilename.startswith(prefix) and subtitleFilename.endswith(
+            subtitleFilePath = os.path.join(searchDirectory, subtitleFilename)
-                "." + MediaDescriptor.SUBTITLE_FILE_EXTENSION
+            subtitleFilenameStem, subtitleFilenameExtension = os.path.splitext(
+                subtitleFilename
+            )
+            if (
+                os.path.isfile(subtitleFilePath)
+                and subtitleFilenameStem.startswith(prefix + '_')
+                and subtitleFilenameExtension.lower() == '.' + normalizedExtension
             ):
+                subtitleFilenames.append(subtitleFilename)
 
-                sesld_result = sesld_match.search(subtitleFilename)
+        expectedSubtitleTrackIndices = {
-                sld_result = None if not sesld_result is None else sld_match.search(subtitleFilename)
+            subtitleTrack.getIndex()
-                
+            for subtitleTrack in self.getSubtitleTracks()
-                if not sesld_result is None:
+        }
+        if strict and len(subtitleFilenames) > len(expectedSubtitleTrackIndices):
+            raise ValueError(
+                f"Found {len(subtitleFilenames)} matching .{normalizedExtension} files "
+                + f"for {len(expectedSubtitleTrackIndices)} subtitle tracks."
+            )
 
-                    subtitleFilePath = os.path.join(searchDirectory, subtitleFilename)
+        for subtitleFilename in subtitleFilenames:
-                    if os.path.isfile(subtitleFilePath):
+            subtitleFilenameStem = os.path.splitext(subtitleFilename)[0]
+            sesld_result = (
+                None
+                if strict
+                else sesld_match.fullmatch(subtitleFilenameStem)
+            )
+            sld_result = (
+                None
+                if sesld_result is not None
+                else sld_match.fullmatch(subtitleFilenameStem)
+            )
 
-                        subtitleFileDescriptor = {}
+            if strict and sesld_result is None and sld_result is None:
-                        subtitleFileDescriptor["path"] = subtitleFilePath
+                raise ValueError(
-                        subtitleFileDescriptor["season"] = int(sesld_result.group(1))
+                    f"Subtitle filename does not match the expected pattern: "
-                        subtitleFileDescriptor["episode"] = int(sesld_result.group(2))
+                    + subtitleFilename
-                        subtitleFileDescriptor["index"] = int(sesld_result.group(3))
+                )
-                        subtitleFileDescriptor["language"] = sesld_result.group(4)
 
-                        dispSet = set()
+            if sesld_result is not None:
-                        dispCaptGroups = sesld_result.groups()
-                        numCaptGroups = len(dispCaptGroups)
-                        if numCaptGroups > 4:
-                            for groupIndex in range(numCaptGroups - 4):
-                                disp = TrackDisposition.fromIndicator(dispCaptGroups[groupIndex + 4])
-                                if disp is not None:
-                                    dispSet.add(disp)
-                        subtitleFileDescriptor["disposition_set"] = dispSet
 
-                        subtitleFileDescriptors.append(subtitleFileDescriptor)
+                subtitleFilePath = os.path.join(searchDirectory, subtitleFilename)
 
-                if not sld_result is None:
+                subtitleFileDescriptor = {}
+                subtitleFileDescriptor["path"] = subtitleFilePath
+                subtitleFileDescriptor["season"] = int(sesld_result.group(1))
+                subtitleFileDescriptor["episode"] = int(sesld_result.group(2))
+                subtitleFileDescriptor["index"] = int(sesld_result.group(3))
+                subtitleFileDescriptor["language"] = sesld_result.group(4)
 
-                    subtitleFilePath = os.path.join(searchDirectory, subtitleFilename)
+                dispSet = set()
-                    if os.path.isfile(subtitleFilePath):
+                dispCaptGroups = sesld_result.groups()
+                numCaptGroups = len(dispCaptGroups)
+                if numCaptGroups > 4:
+                    for groupIndex in range(numCaptGroups - 4):
+                        disp = TrackDisposition.fromIndicator(
+                            dispCaptGroups[groupIndex + 4]
+                        )
+                        if disp is not None:
+                            dispSet.add(disp)
+                subtitleFileDescriptor["disposition_set"] = dispSet
 
-                        subtitleFileDescriptor = {}
+                subtitleFileDescriptors.append(subtitleFileDescriptor)
-                        subtitleFileDescriptor["path"] = subtitleFilePath
-                        subtitleFileDescriptor["index"] = int(sld_result.group(1))
-                        subtitleFileDescriptor["language"] = sld_result.group(2)
 
-                        dispSet = set()
+            if sld_result is not None:
-                        dispCaptGroups = sld_result.groups()
-                        numCaptGroups = len(dispCaptGroups)
-                        if numCaptGroups > 2:
-                            for groupIndex in range(numCaptGroups - 2):
-                                disp = TrackDisposition.fromIndicator(dispCaptGroups[groupIndex + 2])
-                                if disp is not None:
-                                    dispSet.add(disp)
-                        subtitleFileDescriptor["disposition_set"] = dispSet
 
-                        subtitleFileDescriptors.append(subtitleFileDescriptor)
+                subtitleFilePath = os.path.join(searchDirectory, subtitleFilename)
+
+                subtitleFileDescriptor = {}
+                subtitleFileDescriptor["path"] = subtitleFilePath
+                subtitleFileDescriptor["index"] = int(sld_result.group(1))
+                subtitleFileDescriptor["language"] = sld_result.group(2)
+
+                dispSet = set()
+                dispCaptGroups = sld_result.groups()
+                numCaptGroups = len(dispCaptGroups)
+                if numCaptGroups > 2:
+                    for groupIndex in range(numCaptGroups - 2):
+                        disp = TrackDisposition.fromIndicator(
+                            dispCaptGroups[groupIndex + 2]
+                        )
+                        if disp is not None:
+                            dispSet.add(disp)
+                subtitleFileDescriptor["disposition_set"] = dispSet
+
+                subtitleFileDescriptors.append(subtitleFileDescriptor)
+
+        if strict:
+            discoveredTrackIndices = [
+                descriptor['index'] for descriptor in subtitleFileDescriptors
+            ]
+            duplicateTrackIndices = sorted(
+                {
+                    trackIndex
+                    for trackIndex in discoveredTrackIndices
+                    if discoveredTrackIndices.count(trackIndex) > 1
+                }
+            )
+            if duplicateTrackIndices:
+                duplicateDescription = ', '.join(
+                    f"#{index}" for index in duplicateTrackIndices
+                )
+                raise ValueError(
+                    "Multiple external subtitle files refer to subtitle track(s) "
+                    + duplicateDescription
+                    + "."
+                )
+
+            unexpectedTrackIndices = sorted(
+                set(discoveredTrackIndices) - expectedSubtitleTrackIndices
+            )
+            if unexpectedTrackIndices:
+                unexpectedDescription = ', '.join(
+                    f"#{index}" for index in unexpectedTrackIndices
+                )
+                expectedDescription = ', '.join(
+                    f"#{index}" for index in sorted(expectedSubtitleTrackIndices)
+                ) or 'none'
+                raise ValueError(
+                    "External subtitle track index pattern does not match the media "
+                    + f"subtitle tracks: found {unexpectedDescription}; "
+                    + f"expected a subset of {expectedDescription}."
+                )
 
 
         self.__logger.debug(f"searchSubtitleFiles(): Available subtitle files {subtitleFileDescriptors}")
@@ -598,12 +688,19 @@ class MediaDescriptor:
         season: int = -1,
         episode: int = -1,
         preserve_dispositions: bool = False,
+        extension: str = SUBTITLE_FILE_EXTENSION,
+        strict: bool = False,
     ):
         
         # click.echo(f"Season: {season} Episode: {episode}")
         self.__logger.debug(f"importSubtitles(): Season: {season} Episode: {episode}")
 
-        availableFileSubtitleDescriptors = self.searchSubtitleFiles(searchDirectory, prefix)
+        availableFileSubtitleDescriptors = self.searchSubtitleFiles(
+            searchDirectory,
+            prefix,
+            extension=extension,
+            strict=strict,
+        )
         
         self.__logger.debug(f"importSubtitles(): availableFileSubtitleDescriptors: {availableFileSubtitleDescriptors}")
 
@@ -616,7 +713,8 @@ class MediaDescriptor:
                 [
                     d
                     for d in availableFileSubtitleDescriptors
-                    if ((season == -1 and episode == -1)
+                    if (strict
+                        or (season == -1 and episode == -1)
                         or (
                             d.get("season") == int(season)
                             and d.get("episode") == int(episode)
@@ -630,6 +728,7 @@ class MediaDescriptor:
 
         self.__logger.debug(f"importSubtitles(): matchingSubtitleFileDescriptors: {matchingSubtitleFileDescriptors}")
 
+        importedTrackIndices = []
         for msfd in matchingSubtitleFileDescriptors:
             matchingSubtitleTrackDescriptor = [s for s in subtitleTracks if s.getIndex() == msfd["index"]]
             if matchingSubtitleTrackDescriptor:
@@ -643,6 +742,19 @@ class MediaDescriptor:
                 matchingTrack.getTags()["language"] = msfd["language"]
                 if msfd["disposition_set"] and not preserve_dispositions:
                     matchingTrack.setDispositionSet(msfd["disposition_set"])
+                importedTrackIndices.append(matchingTrack.getIndex())
+
+        expectedTrackIndices = sorted(
+            subtitleTrack.getIndex() for subtitleTrack in subtitleTracks
+        )
+        importedTrackIndices = sorted(set(importedTrackIndices))
+        return {
+            "candidate_count": len(availableFileSubtitleDescriptors),
+            "imported_track_indices": importedTrackIndices,
+            "missing_track_indices": sorted(
+                set(expectedTrackIndices) - set(importedTrackIndices)
+            ),
+        }
 
 
     def getConfiguration(self, label: str = ''):

tests/integration/subtrack_mapping/test_cli_bundle.py

@@ -421,6 +421,59 @@ class SubtrackMappingBundleTests(unittest.TestCase):
         self.assertIn("external subtitle payload", extracted_subtitle)
         self.assertNotIn("embedded subtitle payload", extracted_subtitle)
 
+    def test_subtitle_directory_without_prefix_uses_source_basename(self):
+        source_filename = "basename_substitute.mkv"
+        subtitle_directory = self.workdir / "sidecars"
+        subtitle_directory.mkdir()
+        source_path = create_source_fixture(
+            self.workdir,
+            source_filename,
+            [
+                SourceTrackSpec(TrackType.VIDEO, identity="video-0"),
+                SourceTrackSpec(TrackType.AUDIO, identity="audio-1", language="eng"),
+                SourceTrackSpec(
+                    TrackType.SUBTITLE,
+                    identity="embedded-subtitle",
+                    language="eng",
+                    subtitle_lines=("embedded subtitle payload",),
+                ),
+            ],
+        )
+        write_vtt(
+            subtitle_directory / "basename_substitute_2_deu_DEF.vtt",
+            ("external subtitle payload",),
+        )
+
+        completed = run_ffx_convert(
+            self.workdir,
+            self.home_dir,
+            self.database_path,
+            "--video-encoder",
+            "copy",
+            "--no-pattern",
+            "--no-tmdb",
+            "--no-prompt",
+            "--no-signature",
+            "--subtitle-directory",
+            str(subtitle_directory),
+            str(source_path),
+        )
+        self.assertCompleted(completed)
+        self.assertIn("matched subtitle tracks #2", completed.stdout)
+        self.assertIn("Substituting subtitle stream #2", completed.stdout)
+
+        output_path = expected_output_path(self.workdir, source_filename)
+        subtitle_stream = [
+            stream
+            for stream in ffprobe_json(output_path)["streams"]
+            if stream["codec_type"] == "subtitle"
+        ][0]
+        self.assertEqual("deu", get_tag(subtitle_stream, "language"))
+
+        extracted_subtitle = extract_first_subtitle_text(self.workdir, output_path)
+        self.assertIn("external subtitle payload", extracted_subtitle)
+        self.assertNotIn("embedded subtitle payload", extracted_subtitle)
+
     def test_subtitle_prefix_uses_configured_base_directory_when_directory_is_omitted(self):
         source_filename = "substitute_default_s01e01.mkv"
         subtitle_prefix = "substitute_default"

tests/integration/test_cli_unmux.py

@@ -35,7 +35,13 @@ if pytest is not None:
 SRC_ROOT = Path(__file__).resolve().parents[2] / "src"
 
 
-def run_ffx_unmux(workdir: Path, home_dir: Path, database_path: Path, *args: str) -> subprocess.CompletedProcess[str]:
+def run_ffx_unmux(
+    workdir: Path,
+    home_dir: Path,
+    database_path: Path,
+    *args: str,
+    input_text: str | None = None,
+) -> subprocess.CompletedProcess[str]:
     env = os.environ.copy()
     env["HOME"] = str(home_dir)
     existing_pythonpath = env.get("PYTHONPATH", "")
@@ -50,7 +56,14 @@ def run_ffx_unmux(workdir: Path, home_dir: Path, database_path: Path, *args: str
         "unmux",
         *args,
     ]
-    return subprocess.run(command, cwd=workdir, env=env, capture_output=True, text=True)
+    return subprocess.run(
+        command,
+        cwd=workdir,
+        env=env,
+        capture_output=True,
+        input=input_text,
+        text=True,
+    )
 
 
 class UnmuxCliTests(unittest.TestCase):
@@ -164,6 +177,7 @@ class UnmuxCliTests(unittest.TestCase):
             "--label",
             "dball",
             str(source_path),
+            input_text="y\n",
         )
         self.assertCompleted(completed)
 

tests/prepare.sh

@@ -0,0 +1,471 @@
+#!/usr/bin/env bash
+
+set -u
+
+SCRIPT_DIR="$(cd -- "$(dirname -- "${BASH_SOURCE[0]}")" && pwd)"
+ROOT_DIR="$(cd -- "${SCRIPT_DIR}/.." && pwd)"
+VENV_DIR="${FFX_TEST_VENV_DIR:-${ROOT_DIR}/.venv}"
+VENV_BIN_DIR="${VENV_DIR}/bin"
+VENV_PYTHON="${VENV_BIN_DIR}/python"
+VENV_PIP="${VENV_BIN_DIR}/pip"
+
+CHECK_ONLY=0
+READINESS_FAILURES=0
+INSTALL_FAILURES=0
+
+MISSING_REQUIRED_SYSTEM=()
+
+COLOR_RESET=""
+COLOR_GREEN=""
+COLOR_YELLOW=""
+COLOR_RED=""
+
+if [ -t 1 ]; then
+    COLOR_RESET="$(printf '\033[0m')"
+    COLOR_GREEN="$(printf '\033[32m')"
+    COLOR_YELLOW="$(printf '\033[33m')"
+    COLOR_RED="$(printf '\033[31m')"
+fi
+
+usage() {
+    cat <<EOF
+Usage: $(basename "$0") [--check] [--help]
+
+Prepare the repo-local FFX test environment at:
+  ${VENV_DIR}
+
+Actions:
+  - verify or install required system commands for tests
+  - create or reuse the repo-local test virtualenv
+  - install this repository into the venv with Python test and docs extras
+
+Options:
+  --check  Report readiness only. Do not create, install, or modify.
+  --help   Show this help text.
+
+Environment overrides:
+  FFX_TEST_VENV_DIR  Override the test virtualenv path. Defaults to ${ROOT_DIR}/.venv.
+
+Notes:
+  - This script prepares a project-local test environment, not the persistent user bundle.
+  - The persistent bundle setup remains owned by tools/setup.sh.
+EOF
+}
+
+status_ok() {
+    printf '%sok%s' "${COLOR_GREEN}" "${COLOR_RESET}"
+}
+
+status_warn() {
+    printf '%swarn%s' "${COLOR_YELLOW}" "${COLOR_RESET}"
+}
+
+status_fail() {
+    printf '%sfailed%s' "${COLOR_RED}" "${COLOR_RESET}"
+}
+
+report_component() {
+    local level="$1"
+    local label="$2"
+    local detail="$3"
+    local rendered_status=""
+
+    case "${level}" in
+        ok)
+            rendered_status="$(status_ok)"
+            ;;
+        warn)
+            rendered_status="$(status_warn)"
+            ;;
+        *)
+            rendered_status="$(status_fail)"
+            ;;
+    esac
+
+    printf '[%s] %s%s\n' "${rendered_status}" "${label}" "${detail:+: $detail}"
+}
+
+command_exists() {
+    command -v "$1" >/dev/null 2>&1
+}
+
+check_python_venv_support() {
+    python3 -m venv --help >/dev/null 2>&1
+}
+
+check_system_command() {
+    command_exists "$1"
+}
+
+check_venv_python() {
+    [ -x "${VENV_PYTHON}" ]
+}
+
+check_venv_pip() {
+    check_venv_python && "${VENV_PIP}" --version >/dev/null 2>&1
+}
+
+check_venv_ffx() {
+    check_venv_python && "${VENV_PYTHON}" -m ffx version >/dev/null 2>&1
+}
+
+check_venv_pytest() {
+    check_venv_python && "${VENV_PYTHON}" -m pytest --version >/dev/null 2>&1
+}
+
+check_venv_sphinx() {
+    check_venv_python && "${VENV_BIN_DIR}/sphinx-build" --version >/dev/null 2>&1
+}
+
+check_venv_docs_packages() {
+    check_venv_python && "${VENV_PYTHON}" - <<'PY' >/dev/null 2>&1
+import esbonio
+import sphinx
+import sphinx_rtd_theme
+PY
+}
+
+check_editable_install() {
+    check_venv_python && FFX_REPO_ROOT="${ROOT_DIR}" "${VENV_PYTHON}" - <<'PY' >/dev/null 2>&1
+from __future__ import annotations
+
+import os
+from pathlib import Path
+import ffx
+
+repo_root = Path(os.environ["FFX_REPO_ROOT"]).resolve()
+package_path = Path(ffx.__file__).resolve()
+
+raise SystemExit(0 if repo_root in package_path.parents else 1)
+PY
+}
+
+check_python_environment_ready() {
+    check_venv_python &&
+        check_venv_pip &&
+        check_venv_pytest &&
+        check_venv_sphinx &&
+        check_venv_docs_packages &&
+        check_venv_ffx &&
+        check_editable_install
+}
+
+command_detail() {
+    command -v "$1" || printf "command '%s' not found" "$1"
+}
+
+python_venv_detail() {
+    if check_python_venv_support; then
+        printf 'python3 -m venv is available'
+    else
+        printf 'python3 venv support is unavailable'
+    fi
+}
+
+venv_python_detail() {
+    if check_venv_python; then
+        printf '%s' "${VENV_PYTHON}"
+    else
+        printf 'missing %s' "${VENV_PYTHON}"
+    fi
+}
+
+venv_pip_detail() {
+    if check_venv_pip; then
+        "${VENV_PIP}" --version
+    else
+        printf 'missing pip in %s' "${VENV_DIR}"
+    fi
+}
+
+venv_ffx_detail() {
+    if check_venv_ffx; then
+        printf 'ffx import and CLI entry are available'
+    else
+        printf 'ffx is not installed in %s' "${VENV_DIR}"
+    fi
+}
+
+venv_pytest_detail() {
+    if check_venv_pytest; then
+        "${VENV_PYTHON}" -m pytest --version 2>/dev/null | head -n 1
+    else
+        printf 'pytest is not installed in %s' "${VENV_DIR}"
+    fi
+}
+
+venv_sphinx_detail() {
+    if check_venv_sphinx; then
+        "${VENV_BIN_DIR}/sphinx-build" --version 2>&1
+    else
+        printf 'sphinx-build is not installed in %s' "${VENV_DIR}"
+    fi
+}
+
+venv_docs_packages_detail() {
+    if check_venv_docs_packages; then
+        printf 'Sphinx, Read the Docs theme, and Esbonio packages are importable'
+    else
+        printf 'one or more docs packages are missing in %s' "${VENV_DIR}"
+    fi
+}
+
+editable_install_detail() {
+    if check_editable_install; then
+        printf 'ffx resolves from %s' "${ROOT_DIR}"
+    else
+        printf 'ffx does not resolve from the project source tree'
+    fi
+}
+
+report_required_command() {
+    local label="$1"
+    local command_name="$2"
+
+    if check_system_command "${command_name}"; then
+        report_component ok "${label}" "$(command_detail "${command_name}")"
+    else
+        report_component failed "${label}" "$(command_detail "${command_name}")"
+        MISSING_REQUIRED_SYSTEM+=("${command_name}")
+        READINESS_FAILURES=$((READINESS_FAILURES + 1))
+    fi
+}
+
+print_system_status() {
+    MISSING_REQUIRED_SYSTEM=()
+
+    echo "System toolchain status:"
+    report_required_command "git" "git"
+    report_required_command "python3" "python3"
+
+    if check_system_command "python3" && check_python_venv_support; then
+        report_component ok "python3 venv" "$(python_venv_detail)"
+    else
+        report_component failed "python3 venv" "$(python_venv_detail)"
+        MISSING_REQUIRED_SYSTEM+=("python3-venv")
+        READINESS_FAILURES=$((READINESS_FAILURES + 1))
+    fi
+
+    report_required_command "ffmpeg" "ffmpeg"
+    report_required_command "ffprobe" "ffprobe"
+    report_required_command "cpulimit" "cpulimit"
+}
+
+print_python_status() {
+    echo "Repo test and docs virtualenv status:"
+
+    if check_venv_python; then
+        report_component ok "test virtualenv" "$(venv_python_detail)"
+    else
+        report_component failed "test virtualenv" "$(venv_python_detail)"
+        READINESS_FAILURES=$((READINESS_FAILURES + 1))
+    fi
+
+    if check_venv_pip; then
+        report_component ok "test pip" "$(venv_pip_detail)"
+    else
+        report_component failed "test pip" "$(venv_pip_detail)"
+        READINESS_FAILURES=$((READINESS_FAILURES + 1))
+    fi
+
+    if check_venv_pytest; then
+        report_component ok "test pytest" "$(venv_pytest_detail)"
+    else
+        report_component failed "test pytest" "$(venv_pytest_detail)"
+        READINESS_FAILURES=$((READINESS_FAILURES + 1))
+    fi
+
+    if check_venv_sphinx; then
+        report_component ok "docs sphinx" "$(venv_sphinx_detail)"
+    else
+        report_component failed "docs sphinx" "$(venv_sphinx_detail)"
+        READINESS_FAILURES=$((READINESS_FAILURES + 1))
+    fi
+
+    if check_venv_docs_packages; then
+        report_component ok "docs packages" "$(venv_docs_packages_detail)"
+    else
+        report_component failed "docs packages" "$(venv_docs_packages_detail)"
+        READINESS_FAILURES=$((READINESS_FAILURES + 1))
+    fi
+
+    if check_venv_ffx; then
+        report_component ok "test ffx" "$(venv_ffx_detail)"
+    else
+        report_component failed "test ffx" "$(venv_ffx_detail)"
+        READINESS_FAILURES=$((READINESS_FAILURES + 1))
+    fi
+
+    if check_editable_install; then
+        report_component ok "editable source" "$(editable_install_detail)"
+    else
+        report_component failed "editable source" "$(editable_install_detail)"
+        READINESS_FAILURES=$((READINESS_FAILURES + 1))
+    fi
+}
+
+print_status_report() {
+    READINESS_FAILURES=0
+
+    print_system_status
+    echo
+    print_python_status
+}
+
+detect_package_manager() {
+    if command_exists apt-get; then
+        printf 'apt-get\n'
+        return 0
+    fi
+    if command_exists pacman; then
+        printf 'pacman\n'
+        return 0
+    fi
+    return 1
+}
+
+run_root_command() {
+    if [ "${EUID}" -eq 0 ]; then
+        "$@"
+    elif command_exists sudo; then
+        sudo -n "$@"
+    else
+        return 1
+    fi
+}
+
+install_system_requirements() {
+    local package_manager
+
+    if [ "${#MISSING_REQUIRED_SYSTEM[@]}" -eq 0 ]; then
+        return 0
+    fi
+
+    if ! package_manager="$(detect_package_manager)"; then
+        printf 'No supported package manager found for automatic system preparation.\n' >&2
+        INSTALL_FAILURES=$((INSTALL_FAILURES + 1))
+        return 1
+    fi
+
+    case "${package_manager}" in
+        apt-get)
+            printf 'Installing required system dependencies via apt-get...\n'
+            if ! run_root_command apt-get update; then
+                printf 'apt-get update failed or requires interactive sudo.\n' >&2
+                INSTALL_FAILURES=$((INSTALL_FAILURES + 1))
+                return 1
+            fi
+            if ! run_root_command apt-get install -y git python3 python3-venv ffmpeg cpulimit; then
+                printf 'apt-get install failed or requires interactive sudo.\n' >&2
+                INSTALL_FAILURES=$((INSTALL_FAILURES + 1))
+                return 1
+            fi
+            ;;
+        pacman)
+            printf 'Installing required system dependencies via pacman...\n'
+            if ! run_root_command pacman -Sy --noconfirm git python ffmpeg cpulimit; then
+                printf 'pacman install failed or requires interactive sudo.\n' >&2
+                INSTALL_FAILURES=$((INSTALL_FAILURES + 1))
+                return 1
+            fi
+            ;;
+    esac
+
+    return 0
+}
+
+ensure_test_venv() {
+    if ! check_venv_python; then
+        printf 'Creating repo test virtualenv at %s...\n' "${VENV_DIR}"
+        if ! python3 -m venv "${VENV_DIR}"; then
+            printf 'Failed to create test virtualenv at %s.\n' "${VENV_DIR}" >&2
+            INSTALL_FAILURES=$((INSTALL_FAILURES + 1))
+            return 1
+        fi
+    fi
+
+    if ! check_venv_pip; then
+        printf 'Missing pip in %s.\n' "${VENV_DIR}" >&2
+        INSTALL_FAILURES=$((INSTALL_FAILURES + 1))
+        return 1
+    fi
+
+    printf 'Installing FFX package with test and docs extras into %s...\n' "${VENV_DIR}"
+    if ! (
+        cd "${ROOT_DIR}" &&
+        "${VENV_PIP}" install --editable '.[test,docs]'
+    ); then
+        printf 'Failed to install FFX package with test and docs extras into %s.\n' "${VENV_DIR}" >&2
+        INSTALL_FAILURES=$((INSTALL_FAILURES + 1))
+        return 1
+    fi
+
+    return 0
+}
+
+parse_args() {
+    while [ "$#" -gt 0 ]; do
+        case "$1" in
+            --check)
+                CHECK_ONLY=1
+                ;;
+            --help|-h)
+                usage
+                exit 0
+                ;;
+            *)
+                printf 'Unknown option: %s\n\n' "$1" >&2
+                usage >&2
+                exit 2
+                ;;
+        esac
+        shift
+    done
+}
+
+main() {
+    parse_args "$@"
+
+    print_status_report
+
+    if [ "${CHECK_ONLY}" -eq 0 ]; then
+        if [ "${#MISSING_REQUIRED_SYSTEM[@]}" -gt 0 ]; then
+            echo
+            install_system_requirements
+        fi
+
+        if check_python_environment_ready; then
+            echo
+            report_component ok "Python package install" "repo test and docs virtualenv is already ready"
+        elif check_system_command "python3" && check_python_venv_support; then
+            echo
+            ensure_test_venv
+        fi
+
+        echo
+        print_status_report
+    fi
+
+    echo
+    if [ "${INSTALL_FAILURES}" -gt 0 ]; then
+        echo "One or more test preparation steps failed; see the status checks above." >&2
+        exit 1
+    fi
+
+    if [ "${READINESS_FAILURES}" -gt 0 ]; then
+        if [ "${CHECK_ONLY}" -eq 1 ]; then
+            echo "The FFX test and docs environment is incomplete." >&2
+        else
+            echo "Required test or docs components are still missing after preparation." >&2
+        fi
+        exit 1
+    fi
+
+    if [ "${CHECK_ONLY}" -eq 1 ]; then
+        echo "The FFX test and docs environment is ready."
+    else
+        echo "The FFX test and docs environment is prepared."
+    fi
+}
+
+main "$@"

tests/unit/test_cli_subtitle_directory.py

@@ -6,7 +6,9 @@ from pathlib import Path
 import sys
 import tempfile
 import unittest
+from unittest.mock import patch
 
+import click
 from click.testing import CliRunner
 
 
@@ -17,6 +19,10 @@ if str(SRC_ROOT) not in sys.path:
 
 
 from ffx import cli  # noqa: E402
+from ffx.logging_utils import get_ffx_logger  # noqa: E402
+from ffx.media_descriptor import MediaDescriptor  # noqa: E402
+from ffx.track_descriptor import TrackDescriptor  # noqa: E402
+from ffx.track_type import TrackType  # noqa: E402
 
 
 class SubtitleDirectoryCliTests(unittest.TestCase):
@@ -48,6 +54,35 @@ class SubtitleDirectoryCliTests(unittest.TestCase):
             env={**os.environ, "HOME": str(self.home_dir)},
         )
 
+    def make_subtitle_descriptor(self, indices=(2, 3, 4)) -> MediaDescriptor:
+        return MediaDescriptor(
+            context={"logger": get_ffx_logger()},
+            track_descriptors=[
+                TrackDescriptor(
+                    index=index,
+                    source_index=index,
+                    sub_index=subIndex,
+                    track_type=TrackType.SUBTITLE,
+                )
+                for subIndex, index in enumerate(indices)
+            ],
+        )
+
+    def make_import_context(
+        self,
+        subtitleDirectory: Path,
+        noPrompt: bool,
+        yes: bool = False,
+    ) -> dict:
+        return {
+            "subtitle_match_source_basename": True,
+            "subtitle_directory": str(subtitleDirectory),
+            "subtitle_prefix": "",
+            "subtitle_extension": "vtt",
+            "no_prompt": noPrompt,
+            "yes": yes,
+        }
+
     def test_subtitle_prefix_without_directory_or_default_fails(self):
         result = self.invoke_convert("--subtitle-prefix", "dball")
 
@@ -79,6 +114,143 @@ class SubtitleDirectoryCliTests(unittest.TestCase):
 
         self.assertEqual(0, result.exit_code, result.output)
 
+    def test_explicit_directory_without_prefix_enables_basename_matching(self):
+        explicitSubtitleDirectory = self.home_dir / "manual-subtitles"
+        explicitSubtitleDirectory.mkdir(parents=True, exist_ok=True)
+
+        enabled, directory, prefix, matchBasename = cli.resolveSubtitleImportOptions(
+            {},
+            str(explicitSubtitleDirectory),
+            "",
+        )
+
+        self.assertTrue(enabled)
+        self.assertEqual(str(explicitSubtitleDirectory), directory)
+        self.assertEqual("", prefix)
+        self.assertTrue(matchBasename)
+
+    def test_subtitle_extension_accepts_optional_leading_dot(self):
+        self.assertEqual("mkv", cli.normalizeSubtitleExtension(None, None, "mkv"))
+        self.assertEqual("mkv", cli.normalizeSubtitleExtension(None, None, ".mkv"))
+
+    def test_subtitle_extension_rejects_multiple_leading_dots(self):
+        with self.assertRaises(click.BadParameter):
+            cli.normalizeSubtitleExtension(None, None, "..mkv")
+
+    def test_complete_basename_set_does_not_prompt(self):
+        subtitleDirectory = Path(__file__).resolve().parents[1] / "assets" / "subtitles"
+        descriptor = self.make_subtitle_descriptor()
+        context = self.make_import_context(subtitleDirectory, noPrompt=True)
+
+        with patch("ffx.cli.click.confirm") as mockedConfirm:
+            result = cli.importExternalSubtitles(
+                context,
+                descriptor,
+                "A2_t01",
+                -1,
+                -1,
+            )
+
+        self.assertEqual([], result["missing_track_indices"])
+        mockedConfirm.assert_not_called()
+
+    def test_incomplete_basename_set_fails_with_no_prompt(self):
+        descriptor = self.make_subtitle_descriptor()
+        subtitleDirectory = self.home_dir / "partial-subtitles"
+        subtitleDirectory.mkdir()
+        (subtitleDirectory / "episode_2_deu.vtt").write_text(
+            "WEBVTT\n\n",
+            encoding="utf-8",
+        )
+        context = self.make_import_context(subtitleDirectory, noPrompt=True)
+
+        with patch("ffx.cli.click.confirm") as mockedConfirm:
+            with self.assertRaisesRegex(click.ClickException, "--no-prompt is set"):
+                cli.importExternalSubtitles(
+                    context,
+                    descriptor,
+                    "episode",
+                    -1,
+                    -1,
+                )
+
+        mockedConfirm.assert_not_called()
+
+    def test_incomplete_basename_set_can_be_confirmed(self):
+        descriptor = self.make_subtitle_descriptor()
+        subtitleDirectory = self.home_dir / "partial-subtitles"
+        subtitleDirectory.mkdir()
+        (subtitleDirectory / "episode_2_deu.vtt").write_text(
+            "WEBVTT\n\n",
+            encoding="utf-8",
+        )
+        context = self.make_import_context(subtitleDirectory, noPrompt=False)
+
+        with patch("ffx.cli.click.confirm", return_value=True) as mockedConfirm:
+            result = cli.importExternalSubtitles(
+                context,
+                descriptor,
+                "episode",
+                -1,
+                -1,
+            )
+
+        self.assertEqual([3, 4], result["missing_track_indices"])
+        mockedConfirm.assert_called_once()
+
+    def test_incomplete_basename_set_with_yes_does_not_prompt(self):
+        descriptor = self.make_subtitle_descriptor()
+        subtitleDirectory = self.home_dir / "partial-subtitles"
+        subtitleDirectory.mkdir()
+        (subtitleDirectory / "episode_2_deu.vtt").write_text(
+            "WEBVTT\n\n",
+            encoding="utf-8",
+        )
+        context = self.make_import_context(
+            subtitleDirectory,
+            noPrompt=False,
+            yes=True,
+        )
+
+        with patch("ffx.cli.click.confirm") as mockedConfirm:
+            result = cli.importExternalSubtitles(
+                context,
+                descriptor,
+                "episode",
+                -1,
+                -1,
+            )
+
+        self.assertEqual([2], result["imported_track_indices"])
+        self.assertEqual([3, 4], result["missing_track_indices"])
+        mockedConfirm.assert_not_called()
+
+    def test_yes_takes_precedence_over_no_prompt_for_incomplete_set(self):
+        descriptor = self.make_subtitle_descriptor()
+        subtitleDirectory = self.home_dir / "partial-subtitles"
+        subtitleDirectory.mkdir()
+        (subtitleDirectory / "episode_2_deu.vtt").write_text(
+            "WEBVTT\n\n",
+            encoding="utf-8",
+        )
+        context = self.make_import_context(
+            subtitleDirectory,
+            noPrompt=True,
+            yes=True,
+        )
+
+        with patch("ffx.cli.click.confirm") as mockedConfirm:
+            result = cli.importExternalSubtitles(
+                context,
+                descriptor,
+                "episode",
+                -1,
+                -1,
+            )
+
+        self.assertEqual([3, 4], result["missing_track_indices"])
+        mockedConfirm.assert_not_called()
+
 
 if __name__ == "__main__":
     unittest.main()

tests/unit/test_cli_unmux_output_directory.py

@@ -4,6 +4,7 @@ from pathlib import Path
 import sys
 import tempfile
 import unittest
+from unittest.mock import patch
 
 import click
 
@@ -42,7 +43,7 @@ class UnmuxOutputDirectoryTests(unittest.TestCase):
         self.assertEqual(str(Path(tempdir) / "subtitles" / "dball"), resolved_output_directory)
         self.assertTrue(should_create)
 
-    def test_explicit_output_directory_keeps_existing_behavior(self):
+    def test_explicit_output_directory_requires_directory(self):
         with tempfile.TemporaryDirectory() as tempdir:
             context = {
                 "config": StaticConfig(str(Path(tempdir) / "subtitles")),
@@ -57,7 +58,7 @@ class UnmuxOutputDirectoryTests(unittest.TestCase):
             )
 
         self.assertEqual(explicit_output_directory, resolved_output_directory)
-        self.assertFalse(should_create)
+        self.assertTrue(should_create)
 
     def test_subtitles_only_without_label_keeps_existing_behavior(self):
         context = {
@@ -89,6 +90,96 @@ class UnmuxOutputDirectoryTests(unittest.TestCase):
 
         self.assertIn("subtitlesDirectory default", str(caught.exception))
 
+    def test_missing_output_directory_can_be_confirmed_and_created_with_parents(self):
+        with tempfile.TemporaryDirectory() as tempdir:
+            output_directory = Path(tempdir) / "missing" / "parents" / "manual"
+
+            with patch("ffx.cli.click.confirm", return_value=True) as mocked_confirm:
+                created = cli.ensureUnmuxOutputDirectory(
+                    {"dry_run": False},
+                    str(output_directory),
+                )
+
+            self.assertTrue(created)
+            self.assertTrue(output_directory.is_dir())
+            mocked_confirm.assert_called_once()
+
+    def test_tty_carriage_return_accepts_default_directory_creation(self):
+        with tempfile.TemporaryDirectory() as tempdir:
+            output_directory = Path(tempdir) / "missing" / "manual"
+
+            with patch("ffx.cli.sys.stdin.isatty", return_value=True), patch(
+                "ffx.cli.click.getchar",
+                return_value="\r",
+            ) as mocked_getchar, patch("ffx.cli.click.confirm") as mocked_confirm:
+                created = cli.ensureUnmuxOutputDirectory(
+                    {"dry_run": False},
+                    str(output_directory),
+                )
+
+            self.assertTrue(created)
+            self.assertTrue(output_directory.is_dir())
+            mocked_getchar.assert_called_once()
+            mocked_confirm.assert_not_called()
+
+    def test_missing_output_directory_can_be_rejected(self):
+        with tempfile.TemporaryDirectory() as tempdir:
+            output_directory = Path(tempdir) / "missing" / "manual"
+
+            with patch("ffx.cli.click.confirm", return_value=False) as mocked_confirm:
+                with self.assertRaises(click.ClickException) as caught:
+                    cli.ensureUnmuxOutputDirectory(
+                        {"dry_run": False},
+                        str(output_directory),
+                    )
+
+            self.assertFalse(output_directory.exists())
+            self.assertIn("aborted by user", str(caught.exception))
+            mocked_confirm.assert_called_once()
+
+    def test_existing_output_directory_does_not_prompt(self):
+        with tempfile.TemporaryDirectory() as tempdir:
+            output_directory = Path(tempdir) / "manual"
+            output_directory.mkdir()
+
+            with patch("ffx.cli.click.confirm") as mocked_confirm:
+                created = cli.ensureUnmuxOutputDirectory(
+                    {"dry_run": False},
+                    str(output_directory),
+                )
+
+            self.assertFalse(created)
+            mocked_confirm.assert_not_called()
+
+    def test_existing_non_directory_output_path_fails_without_prompt(self):
+        with tempfile.TemporaryDirectory() as tempdir:
+            output_path = Path(tempdir) / "manual"
+            output_path.write_text("not a directory", encoding="utf-8")
+
+            with patch("ffx.cli.click.confirm") as mocked_confirm:
+                with self.assertRaises(click.ClickException) as caught:
+                    cli.ensureUnmuxOutputDirectory(
+                        {"dry_run": False},
+                        str(output_path),
+                    )
+
+            self.assertIn("not a directory", str(caught.exception))
+            mocked_confirm.assert_not_called()
+
+    def test_dry_run_does_not_prompt_or_create_missing_output_directory(self):
+        with tempfile.TemporaryDirectory() as tempdir:
+            output_directory = Path(tempdir) / "missing" / "manual"
+
+            with patch("ffx.cli.click.confirm") as mocked_confirm:
+                created = cli.ensureUnmuxOutputDirectory(
+                    {"dry_run": True},
+                    str(output_directory),
+                )
+
+            self.assertFalse(created)
+            self.assertFalse(output_directory.exists())
+            mocked_confirm.assert_not_called()
+
 
 if __name__ == "__main__":
     unittest.main()

tests/unit/test_media_descriptor_import_subtitles.py

@@ -7,6 +7,7 @@ import unittest
 
 
 SRC_ROOT = Path(__file__).resolve().parents[2] / "src"
+ASSETS_ROOT = Path(__file__).resolve().parents[1] / "assets"
 
 if str(SRC_ROOT) not in sys.path:
     sys.path.insert(0, str(SRC_ROOT))
@@ -20,18 +21,19 @@ from ffx.track_type import TrackType  # noqa: E402
 
 
 class MediaDescriptorImportSubtitlesTests(unittest.TestCase):
-    def make_descriptor(self) -> MediaDescriptor:
+    def make_descriptor(self, indices=(3,)) -> MediaDescriptor:
         return MediaDescriptor(
             context={"logger": get_ffx_logger()},
             track_descriptors=[
                 TrackDescriptor(
-                    index=3,
+                    index=index,
-                    source_index=3,
+                    source_index=index,
-                    sub_index=0,
+                    sub_index=subIndex,
                     track_type=TrackType.SUBTITLE,
                     tags={"language": "eng", "title": "DB Subtitle"},
                     disposition_set={TrackDisposition.DEFAULT},
                 )
+                for subIndex, index in enumerate(indices)
             ],
         )
 
@@ -74,6 +76,110 @@ class MediaDescriptorImportSubtitlesTests(unittest.TestCase):
         self.assertEqual("deu", track.getTags()["language"])
         self.assertEqual({TrackDisposition.FORCED}, track.getDispositionSet())
 
+    def test_strict_basename_import_recognizes_vtt_asset_set(self):
+        descriptor = self.make_descriptor(indices=(2, 3, 4))
+
+        result = descriptor.importSubtitles(
+            str(ASSETS_ROOT / "subtitles"),
+            "A2_t01",
+            strict=True,
+        )
+
+        self.assertEqual(3, result["candidate_count"])
+        self.assertEqual([2, 3, 4], result["imported_track_indices"])
+        self.assertEqual([], result["missing_track_indices"])
+        self.assertEqual(
+            [
+                "A2_t01_2_deu_DEF.vtt",
+                "A2_t01_3_eng.vtt",
+                "A2_t01_4_eng.vtt",
+            ],
+            [
+                Path(track.getExternalSourceFilePath()).name
+                for track in descriptor.getSubtitleTracks()
+            ],
+        )
+
+    def test_strict_basename_import_accepts_dotted_mkv_extension(self):
+        descriptor = self.make_descriptor(indices=(2, 3, 4))
+
+        result = descriptor.importSubtitles(
+            str(ASSETS_ROOT / "subtitles"),
+            "A2_t01",
+            extension=".mkv",
+            strict=True,
+        )
+
+        self.assertEqual(3, result["candidate_count"])
+        self.assertEqual([2, 3, 4], result["imported_track_indices"])
+        self.assertEqual([], result["missing_track_indices"])
+        self.assertTrue(
+            all(
+                track.getExternalSourceFilePath().endswith(".mkv")
+                for track in descriptor.getSubtitleTracks()
+            )
+        )
+
+    def test_strict_basename_import_reports_missing_tracks(self):
+        descriptor = self.make_descriptor(indices=(2, 3, 4))
+
+        with tempfile.TemporaryDirectory() as tmpdir:
+            sidecarPath = Path(tmpdir) / "episode_2_deu.vtt"
+            sidecarPath.write_text("WEBVTT\n\n", encoding="utf-8")
+
+            result = descriptor.importSubtitles(
+                tmpdir,
+                "episode",
+                strict=True,
+            )
+
+        self.assertEqual([2], result["imported_track_indices"])
+        self.assertEqual([3, 4], result["missing_track_indices"])
+
+    def test_strict_basename_import_rejects_too_many_files(self):
+        descriptor = self.make_descriptor(indices=(2,))
+
+        with tempfile.TemporaryDirectory() as tmpdir:
+            for filename in ("episode_2_deu.vtt", "episode_3_eng.vtt"):
+                (Path(tmpdir) / filename).write_text("WEBVTT\n\n", encoding="utf-8")
+
+            with self.assertRaisesRegex(ValueError, "2 matching .* for 1 subtitle tracks"):
+                descriptor.importSubtitles(tmpdir, "episode", strict=True)
+
+    def test_strict_basename_import_rejects_unknown_track_index(self):
+        descriptor = self.make_descriptor(indices=(2, 3, 4))
+
+        with tempfile.TemporaryDirectory() as tmpdir:
+            (Path(tmpdir) / "episode_9_eng.vtt").write_text(
+                "WEBVTT\n\n",
+                encoding="utf-8",
+            )
+
+            with self.assertRaisesRegex(ValueError, "track index pattern does not match"):
+                descriptor.importSubtitles(tmpdir, "episode", strict=True)
+
+    def test_strict_basename_import_rejects_malformed_filtered_filename(self):
+        descriptor = self.make_descriptor(indices=(2, 3, 4))
+
+        with tempfile.TemporaryDirectory() as tmpdir:
+            (Path(tmpdir) / "episode_s01e01_2_deu.vtt").write_text(
+                "WEBVTT\n\n",
+                encoding="utf-8",
+            )
+
+            with self.assertRaisesRegex(ValueError, "expected pattern"):
+                descriptor.importSubtitles(tmpdir, "episode", strict=True)
+
+    def test_strict_basename_import_rejects_duplicate_track_indices(self):
+        descriptor = self.make_descriptor(indices=(2, 3, 4))
+
+        with tempfile.TemporaryDirectory() as tmpdir:
+            for filename in ("episode_2_deu.vtt", "episode_2_eng.vtt"):
+                (Path(tmpdir) / filename).write_text("WEBVTT\n\n", encoding="utf-8")
+
+            with self.assertRaisesRegex(ValueError, "Multiple external subtitle files"):
+                descriptor.importSubtitles(tmpdir, "episode", strict=True)
+
 
 if __name__ == "__main__":
     unittest.main()

tools/merge_dev_into_main.sh

@@ -0,0 +1,448 @@
+#!/usr/bin/env bash
+
+set -euo pipefail
+
+DEV_BRANCH="dev"
+MAIN_BRANCH="main"
+ORIGIN_REMOTE="origin"
+DEFAULT_AGENT_DEVELOPMENT_PATHS=(
+    "AGENTS.md"
+    "SCRATCHPAD.md"
+    "guidance"
+    "requirements"
+    "prompts"
+    "process"
+    "tools/merge_dev_into_main.sh"
+)
+AGENT_DEVELOPMENT_PATHS=("${DEFAULT_AGENT_DEVELOPMENT_PATHS[@]}")
+
+CURRENT_BRANCH="${DEV_BRANCH}"
+ASSUME_YES=0
+DRY_RUN=0
+SKIP_TESTS=0
+
+usage() {
+    cat <<EOF
+Usage: $(basename "$0") [--yes] [--dry-run] [--skip-tests] [--help]
+
+Merge the local ${DEV_BRANCH} branch into ${MAIN_BRANCH}, remove agent-development files
+from ${MAIN_BRANCH}, auto-resolve merge conflicts limited to those cleanup paths,
+create a release merge commit and tag, push to ${ORIGIN_REMOTE}/${MAIN_BRANCH}, and
+switch back to ${DEV_BRANCH}.
+
+Options:
+  --yes         Skip the interactive confirmation prompt.
+  --dry-run     Print the validated release plan without changing git state.
+  --skip-tests  Skip the default pre-release test gate (./tools/test.sh).
+  --help        Show this help text.
+
+Environment overrides:
+  FFX_RELEASE_CLEAN_PATHS  Colon-separated path list to remove from ${MAIN_BRANCH}
+                           after merging ${DEV_BRANCH}. Defaults to:
+                           ${DEFAULT_AGENT_DEVELOPMENT_PATHS[*]}
+EOF
+}
+
+fail() {
+    printf '%s\n' "$*" >&2
+    exit 1
+}
+
+cleanup() {
+    local exit_code="$1"
+
+    trap - EXIT
+
+    if git rev-parse -q --verify MERGE_HEAD >/dev/null 2>&1; then
+        printf 'Merge is incomplete; aborting merge on %s...\n' "${CURRENT_BRANCH}" >&2
+        git merge --abort >/dev/null 2>&1 || true
+    fi
+
+    if [ "${CURRENT_BRANCH}" != "${DEV_BRANCH}" ]; then
+        printf 'Switching back to %s...\n' "${DEV_BRANCH}" >&2
+        git switch "${DEV_BRANCH}" >/dev/null 2>&1 || true
+        CURRENT_BRANCH="${DEV_BRANCH}"
+    fi
+
+    exit "${exit_code}"
+}
+
+load_cleanup_paths() {
+    if [ -n "${FFX_RELEASE_CLEAN_PATHS:-}" ]; then
+        IFS=':' read -r -a AGENT_DEVELOPMENT_PATHS <<< "${FFX_RELEASE_CLEAN_PATHS}"
+    fi
+
+    if [ "${#AGENT_DEVELOPMENT_PATHS[@]}" -eq 0 ]; then
+        fail "Release cleanup path list is empty."
+    fi
+}
+
+path_is_cleanup_target() {
+    local candidate_path="$1"
+    local cleanup_path=""
+
+    for cleanup_path in "${AGENT_DEVELOPMENT_PATHS[@]}"; do
+        case "${candidate_path}" in
+            "${cleanup_path}"|"${cleanup_path}"/*)
+                return 0
+                ;;
+        esac
+    done
+
+    return 1
+}
+
+auto_resolve_cleanup_conflicts() {
+    local unmerged_paths=()
+    local non_cleanup_conflicts=()
+    local remaining_conflicts=()
+    local conflicted_path=""
+
+    mapfile -t unmerged_paths < <(git diff --name-only --diff-filter=U)
+    if [ "${#unmerged_paths[@]}" -eq 0 ]; then
+        return 1
+    fi
+
+    for conflicted_path in "${unmerged_paths[@]}"; do
+        if ! path_is_cleanup_target "${conflicted_path}"; then
+            non_cleanup_conflicts+=("${conflicted_path}")
+        fi
+    done
+
+    if [ "${#non_cleanup_conflicts[@]}" -ne 0 ]; then
+        printf 'Merge produced non-cleanup conflicts:\n' >&2
+        for conflicted_path in "${non_cleanup_conflicts[@]}"; do
+            printf '  - %s\n' "${conflicted_path}" >&2
+        done
+        return 1
+    fi
+
+    printf 'Auto-resolving merge conflicts for release-cleanup paths:\n'
+    for conflicted_path in "${unmerged_paths[@]}"; do
+        printf '  - %s\n' "${conflicted_path}"
+    done
+
+    git rm -r -f --ignore-unmatch "${AGENT_DEVELOPMENT_PATHS[@]}" >/dev/null
+
+    mapfile -t remaining_conflicts < <(git diff --name-only --diff-filter=U)
+    if [ "${#remaining_conflicts[@]}" -ne 0 ]; then
+        printf 'Cleanup conflict auto-resolution left unresolved paths:\n' >&2
+        for conflicted_path in "${remaining_conflicts[@]}"; do
+            printf '  - %s\n' "${conflicted_path}" >&2
+        done
+        return 1
+    fi
+
+    return 0
+}
+
+require_repo_state() {
+    if ! git rev-parse --show-toplevel >/dev/null 2>&1; then
+        fail "This helper must be run inside a git repository."
+    fi
+
+    if ! git show-ref --verify --quiet "refs/heads/${DEV_BRANCH}"; then
+        fail "Local branch '${DEV_BRANCH}' does not exist."
+    fi
+
+    if ! git show-ref --verify --quiet "refs/heads/${MAIN_BRANCH}"; then
+        fail "Local branch '${MAIN_BRANCH}' does not exist."
+    fi
+
+    if ! git remote get-url "${ORIGIN_REMOTE}" >/dev/null 2>&1; then
+        fail "Remote '${ORIGIN_REMOTE}' is not configured."
+    fi
+}
+
+require_dev_checkout() {
+    CURRENT_BRANCH="$(git rev-parse --abbrev-ref HEAD)"
+    if [ "${CURRENT_BRANCH}" != "${DEV_BRANCH}" ]; then
+        fail "Current branch is '${CURRENT_BRANCH}', but '${DEV_BRANCH}' is required."
+    fi
+}
+
+require_clean_worktree() {
+    if [ -n "$(git status --porcelain)" ]; then
+        fail "Local '${DEV_BRANCH}' branch is dirty. Commit, stash, or clean changes first."
+    fi
+}
+
+fetch_remote_state() {
+    printf 'Fetching %s branch and tag state...\n' "${ORIGIN_REMOTE}"
+    git fetch "${ORIGIN_REMOTE}" "${DEV_BRANCH}" "${MAIN_BRANCH}" --tags >/dev/null
+}
+
+branch_divergence_counts() {
+    local branch="$1"
+    local remote_only=""
+    local local_only=""
+
+    if ! git show-ref --verify --quiet "refs/remotes/${ORIGIN_REMOTE}/${branch}"; then
+        fail "Remote branch '${ORIGIN_REMOTE}/${branch}' does not exist."
+    fi
+
+    read -r remote_only local_only < <(
+        git rev-list --left-right --count \
+            "refs/remotes/${ORIGIN_REMOTE}/${branch}...refs/heads/${branch}"
+    )
+
+    printf '%s %s\n' "${remote_only}" "${local_only}"
+}
+
+fast_forward_branch_to_remote() {
+    local branch="$1"
+    local remote_ref="refs/remotes/${ORIGIN_REMOTE}/${branch}"
+    local current_head=""
+
+    current_head="$(git rev-parse --abbrev-ref HEAD)"
+
+    printf "Fast-forwarding local branch '%s' to '%s/%s'...\n" \
+        "${branch}" \
+        "${ORIGIN_REMOTE}" \
+        "${branch}"
+
+    if [ "${current_head}" = "${branch}" ]; then
+        git merge --ff-only "${remote_ref}" >/dev/null
+        return 0
+    fi
+
+    git branch -f "${branch}" "${remote_ref}" >/dev/null
+}
+
+sync_release_source_branch() {
+    local branch="$1"
+    local remote_only=""
+    local local_only=""
+
+    read -r remote_only local_only < <(branch_divergence_counts "${branch}")
+
+    if [ "${remote_only}" -ne 0 ] && [ "${local_only}" -ne 0 ]; then
+        fail "Local branch '${branch}' has diverged from '${ORIGIN_REMOTE}/${branch}' (${local_only} local-only commit(s), ${remote_only} remote-only commit(s)). Reconcile the branches first."
+    fi
+
+    if [ "${remote_only}" -ne 0 ]; then
+        fast_forward_branch_to_remote "${branch}"
+    fi
+
+    if [ "${local_only}" -ne 0 ]; then
+        printf "Notice: local branch '%s' is ahead of '%s/%s' by %s commit(s); release will use the local tip.\n" \
+            "${branch}" \
+            "${ORIGIN_REMOTE}" \
+            "${branch}" \
+            "${local_only}"
+    fi
+}
+
+sync_release_target_branch() {
+    local branch="$1"
+    local remote_only=""
+    local local_only=""
+
+    read -r remote_only local_only < <(branch_divergence_counts "${branch}")
+
+    if [ "${remote_only}" -ne 0 ] && [ "${local_only}" -ne 0 ]; then
+        fail "Local branch '${branch}' has diverged from '${ORIGIN_REMOTE}/${branch}' (${local_only} local-only commit(s), ${remote_only} remote-only commit(s)). Reconcile the branches first."
+    fi
+
+    if [ "${local_only}" -ne 0 ]; then
+        fail "Local branch '${branch}' is ahead of '${ORIGIN_REMOTE}/${branch}' by ${local_only} commit(s). Push or reconcile first so the release starts from the published ${branch} tip."
+    fi
+
+    if [ "${remote_only}" -ne 0 ]; then
+        fast_forward_branch_to_remote "${branch}"
+    fi
+}
+
+resolve_release_version() {
+    local version_from_pyproject=""
+    local version_from_constants=""
+
+    version_from_pyproject="$(
+        sed -n 's/^version = "\(.*\)"$/\1/p' pyproject.toml | head -n 1
+    )"
+    version_from_constants="$(
+        sed -n "s/^VERSION='\(.*\)'$/\1/p" src/ffx/constants.py | head -n 1
+    )"
+
+    if [ -z "${version_from_pyproject}" ]; then
+        fail "Could not resolve release version from pyproject.toml."
+    fi
+
+    if [ -z "${version_from_constants}" ]; then
+        fail "Could not resolve release version from src/ffx/constants.py."
+    fi
+
+    if [ "${version_from_pyproject}" != "${version_from_constants}" ]; then
+        fail "Version mismatch: pyproject.toml=${version_from_pyproject}, src/ffx/constants.py=${version_from_constants}."
+    fi
+
+    printf '%s\n' "${version_from_pyproject}"
+}
+
+require_release_tag_available() {
+    local release_version="$1"
+    local release_tag="v${release_version}"
+
+    if git rev-parse -q --verify "refs/tags/${release_tag}" >/dev/null 2>&1; then
+        fail "Tag '${release_tag}' already exists."
+    fi
+
+    if git rev-parse -q --verify "refs/tags/${release_version}" >/dev/null 2>&1; then
+        fail "Bare tag '${release_version}' already exists; refusing to create ambiguous release tags."
+    fi
+}
+
+run_pre_release_tests() {
+    if [ "${SKIP_TESTS}" -eq 1 ]; then
+        printf 'Skipping pre-release tests.\n'
+        return 0
+    fi
+
+    if [ ! -x "./tools/test.sh" ]; then
+        fail "Missing executable test runner at ./tools/test.sh."
+    fi
+
+    printf 'Running pre-release tests via ./tools/test.sh...\n'
+    ./tools/test.sh
+}
+
+print_release_plan() {
+    local release_version="$1"
+    local release_tag="v${release_version}"
+    local release_commit_message="Release ${release_tag}"
+
+    printf 'Dry run only. Planned steps:\n'
+    printf '1. Ensure current branch is %s and the worktree is clean.\n' "${DEV_BRANCH}"
+    printf '2. Fetch %s, fast-forward local %s and %s from %s when safe, and fail on divergence or unpublished local %s commits.\n' \
+        "${ORIGIN_REMOTE}" \
+        "${DEV_BRANCH}" \
+        "${MAIN_BRANCH}" \
+        "${ORIGIN_REMOTE}" \
+        "${MAIN_BRANCH}"
+    if [ "${SKIP_TESTS}" -eq 1 ]; then
+        printf '3. Skip the pre-release test gate.\n'
+    else
+        printf '3. Run ./tools/test.sh as the pre-release test gate.\n'
+    fi
+    printf '4. Switch to %s and merge %s with --no-ff --no-commit.\n' "${MAIN_BRANCH}" "${DEV_BRANCH}"
+    printf '5. Auto-resolve merge conflicts limited to release-cleanup paths and remove them from %s:\n' "${MAIN_BRANCH}"
+    local cleanup_path=""
+    for cleanup_path in "${AGENT_DEVELOPMENT_PATHS[@]}"; do
+        printf '   - %s\n' "${cleanup_path}"
+    done
+    printf '6. Create merge commit: %s\n' "${release_commit_message}"
+    printf '7. Create annotated tag: %s\n' "${release_tag}"
+    printf '8. Push %s to %s/%s with --follow-tags.\n' "${MAIN_BRANCH}" "${ORIGIN_REMOTE}" "${MAIN_BRANCH}"
+    printf '9. Switch back to %s.\n' "${DEV_BRANCH}"
+}
+
+trap 'cleanup $?' EXIT
+
+while [ "$#" -gt 0 ]; do
+    case "$1" in
+        --yes)
+            ASSUME_YES=1
+            ;;
+        --dry-run)
+            DRY_RUN=1
+            ;;
+        --skip-tests)
+            SKIP_TESTS=1
+            ;;
+        --help|-h)
+            usage
+            exit 0
+            ;;
+        *)
+            usage >&2
+            fail "Unknown option: $1"
+            ;;
+    esac
+    shift
+done
+
+load_cleanup_paths
+require_repo_state
+require_dev_checkout
+require_clean_worktree
+fetch_remote_state
+sync_release_source_branch "${DEV_BRANCH}"
+sync_release_target_branch "${MAIN_BRANCH}"
+
+RELEASE_VERSION="$(resolve_release_version)"
+RELEASE_TAG="v${RELEASE_VERSION}"
+RELEASE_COMMIT_MESSAGE="Release ${RELEASE_TAG}"
+require_release_tag_available "${RELEASE_VERSION}"
+
+printf 'This will merge %s into %s, remove agent-development files on %s,\n' "${DEV_BRANCH}" "${MAIN_BRANCH}" "${MAIN_BRANCH}"
+printf 'auto-resolve cleanup-path conflicts, run the pre-release gate%s, create %s,\n' \
+    "$([ "${SKIP_TESTS}" -eq 1 ] && printf ' (skipped)' || printf '')" \
+    "${RELEASE_TAG}"
+printf 'push to %s/%s, and switch back to %s.\n' \
+    "${ORIGIN_REMOTE}" \
+    "${MAIN_BRANCH}" \
+    "${DEV_BRANCH}"
+
+if [ "${ASSUME_YES}" -ne 1 ]; then
+    printf 'Are you sure? [y/N] '
+    read -r confirmation
+    case "${confirmation}" in
+        y|Y|yes|YES)
+            ;;
+        *)
+            fail "Aborted by user."
+            ;;
+    esac
+fi
+
+if [ "${DRY_RUN}" -eq 1 ]; then
+    print_release_plan "${RELEASE_VERSION}"
+    exit 0
+fi
+
+run_pre_release_tests
+require_clean_worktree
+fetch_remote_state
+sync_release_source_branch "${DEV_BRANCH}"
+sync_release_target_branch "${MAIN_BRANCH}"
+require_release_tag_available "${RELEASE_VERSION}"
+
+git switch "${MAIN_BRANCH}" >/dev/null
+CURRENT_BRANCH="${MAIN_BRANCH}"
+
+printf 'Merging %s into %s...\n' "${DEV_BRANCH}" "${MAIN_BRANCH}"
+if ! git merge --no-ff --no-commit "${DEV_BRANCH}"; then
+    if ! auto_resolve_cleanup_conflicts; then
+        fail "Merge from '${DEV_BRANCH}' into '${MAIN_BRANCH}' failed."
+    fi
+fi
+
+if ! git rev-parse -q --verify MERGE_HEAD >/dev/null 2>&1; then
+    fail "'${MAIN_BRANCH}' is already up to date with '${DEV_BRANCH}'. Nothing to merge."
+fi
+
+printf 'Removing agent-development files from %s...\n' "${MAIN_BRANCH}"
+git rm -r -f --ignore-unmatch "${AGENT_DEVELOPMENT_PATHS[@]}" >/dev/null
+
+if git diff --cached --quiet; then
+    fail "No staged changes are present after merging '${DEV_BRANCH}' into '${MAIN_BRANCH}'."
+fi
+
+printf 'Creating release merge commit: %s\n' "${RELEASE_COMMIT_MESSAGE}"
+git commit -m "${RELEASE_COMMIT_MESSAGE}"
+
+printf 'Creating annotated tag: %s\n' "${RELEASE_TAG}"
+git tag -a "${RELEASE_TAG}" -m "FFX ${RELEASE_VERSION}"
+
+printf 'Pushing %s and annotated tags to %s...\n' "${MAIN_BRANCH}" "${ORIGIN_REMOTE}"
+git push "${ORIGIN_REMOTE}" "${MAIN_BRANCH}" --follow-tags
+
+printf 'Switching back to %s...\n' "${DEV_BRANCH}"
+git switch "${DEV_BRANCH}" >/dev/null
+CURRENT_BRANCH="${DEV_BRANCH}"
+
+printf 'Release merge complete: %s pushed to %s/%s and tagged as %s.\n' \
+    "${RELEASE_COMMIT_MESSAGE}" \
+    "${ORIGIN_REMOTE}" \
+    "${MAIN_BRANCH}" \
+    "${RELEASE_TAG}"