storyboard-understand
Skill2 starsChanged 7 months agoMITView on GitHub
Turn storyboard OCR into an evidence-backed product summary with likely views, capabilities, and open questions.
In this repository17 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- compare-video-runsSkill
- install-skill-packSkill
- layout-safety-reviewSkill
- package-review-promptSkill
- review-bundleSkill
- segment-evidenceSkill
- segment-storyboardSkill
- skill-catalogSkill
- source-media-signalsSkill
- storyboard-extractSkill
- storyboard-ocrSkill
- storyboard-transitionsSkill
- video-artifact-intakeSkill
- video-shotsSkill
- video-technical-reviewSkill
The file
We split this skill into the rules it gives an agent
4 rules. About 0 are common, 0 are distinctive, and 4 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.
4
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 8
Use this after storyboard-ocr when an agent needs a first-pass structured understanding rather than raw OCR lines.
- Very specific: line 28
If storyboard.transitions.json exists in the storyboard directory, the skill can incorporate transition-derived flow segments.
- Very specific: line 40
Run after storyboard-ocr. For stronger flow summaries, run storyboard-transitions first.
- Very specific: line 54
Do not present guesses as facts.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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
Storyboard Understand
Use this after storyboard-ocr when an agent needs a first-pass
structured understanding rather than raw OCR lines.
This skill owns:
- app-name guesses from OCR evidence
- likely role/view extraction
- capability claims with cited lines and frame indices
- explicit open questions where inference is still weak
Inputs
Provide either storyboardDir or ocrPath:
{
"storyboardDir": "./output/storyboard"
}If storyboard.transitions.json exists in the storyboard directory, the skill
can incorporate transition-derived flow segments.
Outputs
The skill writes storyboard.summary.json. The summary includes app-name
guesses, likely views, OCR quality notes, sampling notes, interaction segments,
likely flow, capability claims with frame/line evidence, text-dominance signals,
and open questions.
Sequencing Guidance
Run after storyboard-ocr. For stronger flow summaries, run
storyboard-transitions first. Use this output to orient a reviewer, then
verify important claims against storyboard.ocr.json and the source frame
images.
Interpretation Notes
Capability claims are pattern-based and must cite OCR evidence. App names and
view names are guesses, especially when OCR is sparse or narration text
dominates. textDominance.likelyNarrationDominated means the sampled text may
describe the product rather than come from the UI itself.
Abstention Rules
Do not present guesses as facts. Abstain from final product claims when
ocrQuality.lowSignal is true, most frames are weak/rejected, or the relevant
claim appears only in narration-like text. Keep open questions visible rather
than filling gaps from domain assumptions.
Failure Modes
Missing OCR input fails the run. Missing transitions does not fail the run, but flow and interaction segments may be less specific. Poor sampling or OCR noise can produce generic app/view labels; inspect evidence before reusing them.
Repo-side runner:
node --import tsx scripts/harness/storyboard-understand.ts
Installed-pack runner:
node ./.video-evaluator/agent/run-tool.mjs storyboard-understand
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.

