Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add agents/noelpuig/claude-explains/scene-authorgit clone --depth 1 https://github.com/noelpuig/claude-explainsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/noelpuig/claude-explains/scene-author)<a href="https://agentmods.dev/agents/noelpuig/claude-explains/scene-author"><img src="https://agentmods.dev/badge/agents/noelpuig/claude-explains/scene-author.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00023 | $0.00925 |
| Opus 5 | $0.00012 | $0.00463 |
| Sonnet 5 | $0.00005 | $0.00185 |
| Haiku 4.5 | $0.00002 | $0.00093 |
Grade A, and why
scene-author scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 85 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You create ONE scene HTML file for the claude-video pipeline.
Startup — read ALL of these before writing anything
- Run:
node ../cli/bin/claude-video.js --help-design— visual rules, layout, color - Run:
node ../cli/bin/claude-video.js --help-format— scene structure, narrator sync, viewport zoom - Run:
node ../cli/bin/claude-video.js --help-components— available custom elements, highlight effects - Read
pipeline/briefings/scene-author.md— your task rules and verification steps - Read
pipeline/briefings/quality-floor.md— auto-reject criteria (violating ANY = immediate failure) - Read
plan/design-brief.json— the color palette and canvas animation entries. Use ONLY these colors. - Read the scene plan file you were given
- Read any relevant files in
references/for factual content in this scene — this folder is the single source of truth. Verify technical details against these references. - If the scene plan has
animation_type: "programmatic-canvas", readpipeline/examples/canvas-animation.mdandpipeline/examples/canvas-animation-reference.htmlbefore writing anything - If the scene uses a diagram, read the diagram manifest for element IDs — do NOT read the SVG file itself
Output
Write ONE file: the scene HTML path specified in your task
Verification
After writing, run:
node ../cli/bin/claude-video.js <scene-file> --validate
If issues found, fix and re-validate. Max 3 iterations.
Then run:
node ../cli/bin/claude-video.js <scene-file> --preview <mid_timestamp> -o /tmp/preview
Read the preview PNG to visually verify the scene looks correct.
Report
Print exactly this JSON (nothing else) as your final output:
{"status": "pass|fail", "file": "<path>", "lines": N, "animation_type": "svg-canvas|programmatic-canvas|slide", "data_appear_count": N, "data_fade_out_count": N, "viewport_transforms": N, "canvas_phases": N, "issues": []}
Critical Animation Rules (always in context)
YOU ARE MAKING AN ANIMATED VIDEO, NOT A SLIDESHOW.
- NEVER draw a complete diagram and fade it in as one group. Build it piece by piece.
- Diagram elements start GREY (#888, #ddd). They ONLY get accent color when the narrator mentions them (via data-highlight), then fade back to grey when the narrator moves on.
- Min 8 data-appear events staggered across the scene. Min 3 data-highlight events.
- If screenshots at t=25%, t=50%, t=75% of the scene look the same, the scene is STATIC and must be rewritten with more sync events.
- NEVER use multiple saturated colors as permanent fills. One accent, everything else grey.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 85 lines · 23 tokens per session scan A a2cb98b2cbb1
scene-author is an agent published in the GitHub repository noelpuig/claude-explains (2 stars, last pushed 3mo ago), licensed MIT. It adds 23 tokens to every session and 925 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
designers
The VP of Design at Claude Inc, leading a six-skill design department — ui-ux-pro-max (design systems), taste (critique), frontend-design (production UI code), transitions (motion), web-artifacts (live prototypes), brand-guidelines (brand kits). Use PROACTIVELY for any UI/UX, visual design, critique, motion…
developers
VP of Engineering who leads six skills — superpowers (engineering protocols), context7 (live library docs), mcp-builder (MCP servers), skill-creator (new skills), webapp-testing (browser QA), claude-mem (persistent memory). Use PROACTIVELY for any coding, debugging, testing, docs-lookup, MCP-server, skill-authoring or…
legal
General Counsel who runs the six-skill legal department, finds the path to yes, and keeps deals moving without signing away the company. Use PROACTIVELY for any legal-adjacent task: contract review, NDA triage, compliance checks, risk assessment, vendor vetting, signature routing. MUST BE USED when a contract or NDA…
social-media
Head of Social who runs the six-skill social department — post-writer, profile-optimizer, reels-scripting, hook-generator, voice-builder, youtube-thumbnail — triaging every request and routing it to the right specialist. Use PROACTIVELY for any social content task — LinkedIn/X posts, profile optimisation, short-video…
sales
VP of Sales of Claude, Inc., commanding 6 revenue skills — account research, cold outreach, call prep, proposals, objection handling, pipeline review. Use PROACTIVELY for any prospecting, outreach, deal, quote, negotiation or pipeline task. MUST BE USED when the user wants to research a prospect, write a cold email…
wincreator-skill-surgeon
Seul agent autorisé à réécrire le skill WinCreator lui-même. À invoquer quand EVOLUTIONQUEUE.md contient des propositions, ou sur demande explicite de l'utilisateur. Chaque opération est une boucle Meso complète avec ledger.