movie-commentary
Skill23 starsChanged 41 days agoMITView on GitHub
Turn one fixed local movie into a source-linked narrated commentary cut through a durable project, a validated plan, immutable frozen shards, explicitly approved per-shard host execution and verified full-cut delivery. Use for movie or long-video commentary projects, resuming them, or diagnosing why delivery is refused.
In this repository30 other files it gives its agents.
- AGENTS.mdAGENTS.md
- src/AGENTS.mdAGENTS.md
- tests/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- av-eventsSkill
- creative-concept-designSkill
- educational-explainerSkill
- ffmpeg-productionSkill
- footage-editSkill
- gemini-visualizeSkill
- gr-advisorSkill
- hardware-evidence-captureSkill
- image-generationSkill
- mlflow-tracesSkill
- plugin-maintenanceSkill
- qwen-image-integrationSkill
- qwen-video-integrationSkill
- research-brief-builderSkill
- research-visualization-blenderSkill
- research-visualization-freecadSkill
- reverse-search-video-frameSkill
- spatial-video-analysisSkill
- tts-productionSkill
- video-explainerSkill
- video-generationSkill
- video-productionSkill
- video-researchSkill
- video-to-skillSkill
- video-translationSkill
- weaviate-setupSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 16
Call commentary_inspect first and follow recommended_stage.
- Distinctive: line 40
Run a shard only after an explicit approval from the person or authority responsible for the source media.
- Very specific: line 46
Write <shard_id>.mp4 and exec_report.json at the returned paths.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Movie commentary
The video-explainer server provides seven commentary_* tools. They prepare projects, check plans, freeze
shards, record approvals, assemble the cut and validate delivery. They never write narration, interpret
the film or render a shard, and no tool spawns an agent. The full contract is in
movie-commentary.md in the source package
(../video-research-resources/docs/integrations/movie-commentary.md in the installed support tree).
Route by state
Call commentary_inspect first and follow recommended_stage.
| Stage | Action |
| no project | commentary_prepare with the movie path and its exact SHA256 |
analyze_source | write watch notes under plan/watch_notes/*.md from evidence you actually observed |
author_plan | write plan/editing_plan.json and plan/narration_script.md, run commentary_validate_plan until it is valid, then commentary_freeze_shards |
approve_execute_or_assemble | get explicit approval for each shard, then execute exactly that scope |
| all shards reported | commentary_assemble, then commentary_validate_delivery |
blocked_source_changed | stop; the project is bound to bytes that changed, so start a new project |
Plan
editing_plan.json has schema vrm/movie-commentary-plan/v1 and the project's source_sha256.
Its segments are numbered SEG_0001 onward. Each segment has:
narration.text, repeated verbatim with its ID in the narration script;visual_plan.rough_interval_sec, which must lie withinsource_cut_max_sec;visual_plan.movie_locator.evidence_refs, existing files underplan/watch_notes/only;audio_plan.source_audio_mode:ducked_bedormuted;audio_plan.bgm_mode:none, orlicensedonly when the execution facts name a BGM manifest.
Approval and execution
- Run a shard only after an explicit approval from the person or authority responsible for the source media.
- Pass that approval to
commentary_approve_shardwith the exact shard and source SHA256. - Record the caller's media authority and the exact execution scope. Existing explicit approval persists for those frozen bytes; changed source or scope requires a new approval.
- Execute only the returned scope: the shard manifest, the execution facts and the source. Do not change narration, segment order or creative intent.
- Write
<shard_id>.mp4andexec_report.jsonat the returned paths. - Measured failures stay failures. An unresolved segment or a failed QA check blocks delivery.
Delivery
commentary_assemble concatenates the approved shard MP4s with the pinned ffmpeg, fully decodes the cut and
compares its duration with the summed shard durations. commentary_validate_delivery refuses delivery
when any of these is missing or changed:
- the source, plan, script, evidence or execution facts;
- a frozen shard, an approval or a report;
- a shard MP4, the final MP4 or the final QA.
A blocked project is better than a false pass.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

