agentleFS
mocchaleramocchalera/video-os-v2/.agents/skills/vertical-social-composition/SKILL.md

vertical-social-composition

Skill4 starsChanged 2 months agoMITView on GitHub

Resolve registered vertical composition inputs from source-grounded framing evidence, source A/V identity, layout anchors, and safe degradation. Use when a blueprint contains framing candidates or visual intents for a vertical delivery.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read 04_plan/edit_blueprint.yaml, 04_plan/selects_candidates.yaml, and the referenced framing policy.

    line 12
  2. Very specific:

    When subject-caption collision QA is requested, use a source-grounded subject-occupancy-track/v1 whose track labels are explicitly not …

    line 18
  3. Very specific:

    Run npm run validate:all-local, npm run typecheck, and npm run verify after changing a policy or consuming an artifact.

    line 31

≈ “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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Vertical Social Composition

Inspect and resolve the existing framing policy, reframe candidate, and registered visual intent before any visual treatment is considered. Keep source asset, segment, range, source hash, A/V geometry, and first/representative/last evidence attached to the result.

Inspection order

  1. Read 04_plan/edit_blueprint.yaml, 04_plan/selects_candidates.yaml, and the referenced framing policy. Confirm that policy_refs.composition_policy_ref resolves inside the project and that any source hash is current.
  2. Inspect the registered visual intent and its source_evidence. Verify the existing RFA-008 to RFA-010 candidate when a candidate reference is present. Treat unavailable optional vision adapters as a source-grounded degrade, not as permission to invent a crop.
  3. Resolve vertical_composition_policy_ref through schemas/vertical-composition-policy.schema.json and runtime/visual/vertical-composition.ts. Supply policy-owned person occupancy, headroom, look-room, hands, microphone, evidence, layout-anchor, and zoom-intent inputs.
  4. Attach first, representative, and last frame observations and preserve source A/V identity. Do not force a punch-in for a talking head. Use the policy's safe-degrade mode or leave a human hold when evidence is incomplete.
  5. Review compiled 05_timeline/timeline.json provenance. A vertical-composition-resolution/v1 result is an input receipt; it is not a graphical compositor or Studio edit.

When subject-caption collision QA is requested, use a source-grounded subject-occupancy-track/v1 whose track labels are explicitly not person identity. Resolve thresholds and lower/upper proposal anchors from the same project-contained vertical composition policy, then bind the track, renderer layout snapshot, policy hash, renderer capability, verdict, and review item to one immutable social-review generation. Missing or mixed-generation evidence is an incomplete/human hold. Candidate anchors never authorize caption text, timing, approval, or canonical timeline mutation, and auto-move remains off.

Speech caption text, timing, and approval remain owned by the existing caption runtime and FFmpeg/libass route. Registered graphical content elements remain eligible for the existing Remotion or HyperFrames route in the later visual-treatment milestone; visual treatment patches, platform profiles, and Studio controls remain separate. This skill only resolves their inputs and does not rewrite the canonical timeline by hand.

Checks

Run npm run validate:all-local, npm run typecheck, and npm run verify after changing a policy or consuming an artifact. Inspect 05_timeline/timeline.json and its provenance; keep 04_plan/edit_blueprint.yaml as the authoring source.

Record the resolved policy reference, policy hash, candidate evidence, receipt status, and any human hold in the canonical timeline provenance or an existing review/route receipt. Do not invent a standalone composition artifact. Keep renderer application and Studio graphical controls for the later milestone.

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.