agentleFS
imMamdouhaboammarimMamdouhaboammar/linkedin-animated-infographics/docs/agents.md

linkedin-animated-infographics

AGENTS.md14 starsChanged 3 months agoMITView on GitHub

The active worker inventory is declared in helper/modules.json and the executable order is declared in architecture/plugin-graph.json. Every agent returns a bounded artifact or verdict to the parent workflow. Agents do not coordinate peer agents directly. design-study reports ranked evidence, confidence, provenance/rights, focused contexts, and explicit READY/HOLD/SKIP status. Diagnoses visual references into reusable design DNA before concept generation. It owns reference-dna when references are present and returns build/design-study.json. Defines the factual boundary before creative work begins. It classifies claims, product states,…

The file

We split this agents.md into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is 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.

  • 1

    Very specific

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

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

    Read-only independent acceptance worker.

    line 79

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md 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
agents.md

Agents

The active worker inventory is declared in helper/modules.json and the executable order is declared in architecture/plugin-graph.json. Every agent returns a bounded artifact or verdict to the parent workflow. Agents do not coordinate peer agents directly.

Visual study contract

design-study reports ranked evidence, confidence, provenance/rights, focused contexts, and explicit READY/HOLD/SKIP status.

Complete workflow agents

design-study

Diagnoses visual references into reusable design DNA before concept generation. It owns reference-dna when references are present and returns build/design-study.json.

evidence-checker

Defines the factual boundary before creative work begins. It classifies claims, product states, metrics, named identities, logos, proof, and UI evidence, then returns build/evidence.json. It owns evidence-traceability at intake.

asset-curator

Resolves named AI/tool logos and mascot identities before concept generation. Exact user-supplied official assets have priority; supported named AI/tool identities then resolve through Lobe after reading the current Lobe icon skill instructions. It records exact provenance, requires a local or embedded render copy, blocks generated official lookalikes, and returns build/asset-plan.json.

creative-director

Generates at least three evidence-safe creative directions after identity assets are verified. Each direction contains a visual hook, copy hook, aha mechanic, story relationship, dominant visual anchor, structural archetype, containment strategy, negative-space strategy, motion job, evidence dependencies, risks, and why it earns attention. It returns build/creative-concepts.json.

This agent applies hooked-design-copy, creative-payoff, and clean-creative-structure. It must not use spectacle or generic card grids as a substitute for explanation.

story-architect

Turns the selected creative direction into a deterministic Info-stories brief. It resolves Story Archetype, Visual Style, Story House, Motion Patterns, design dials, and the selected clean-structure requirements, then returns build/story-brief.json.

palette-curator

Resolves the Story House and semantic token roles. It enforces creative-attractive-restrained palette behavior and the contrast-discipline research gate. It returns build/palette-check.json.

type-curator

Chooses typography before copy fitting. Explicit user typography wins when render-safe, followed by supplied or bundled local assets, then a curated deterministic system direction. It records headline/body roles, fallbacks, minimum feed sizes, pairing rationale, and an allowed system, embedded, or local-file loading strategy in build/type-spec.json. Remote render-time font loading is a blocking failure.

copy-compressor

Compresses source material into infographic-sized design copy while protecting facts, names, numbers, mechanisms, voice, the approved copy hook, and the minimum feed-size constraints from the type spec. It applies prose-specificity, voice-preservation, and evidence constraints, then returns build/artboard-copy.json.

layout-composer

Turns the selected concept, story brief, asset plan, type spec, copy, palette, and optional reference study into a static hierarchy and topology specification. It applies center-first-composition, clean-creative-structure, verified-identity-assets, intentional-typography, design-dials, structural-originality, and reference-dna when active. It returns build/layout-spec.json.

caption-writer

Writes the LinkedIn caption and first comment. It applies hooked-design-copy to the opening, keeps one caption archetype, preserves evidence, and returns build/caption.md plus build/first-comment.md.

artboard-builder

Builds the approved 1080x1350 HTML still and rendered PNG. It preserves the selected creative payoff and clean structure, uses only identity-locked assets from the asset plan, implements the approved type spec without remote font loading, preserves Story House tokens, center-first layout or documented exception, UI fidelity, and contrast. It returns build/post.html and build/still.png.

motion-director

Chooses zero to two meaning-driven Motion Patterns after the still is approved. Motion must communicate hierarchy, sequence, reveal, state, route, or the selected creative payoff. It returns build/motion-direction.json.

mascot-animator

Conditional worker for a named or official mascot. It consumes the exact verified local SVG from build/asset-plan.json when the complete workflow resolved the identity. That SVG may be an exact user/task asset or a verified Lobe asset for a supported named AI/tool identity. Direct focused mascot work keeps the exact user-SVG requirement. The worker inspects real geometry, preserves identity, selects a communication-led direction, and returns build/mascot/motion-contract.json. Missing verified identity is a HOLD.

motion-engineer

Implements the approved seekable motion direction on the approved still. It preserves the resolved story brief, frame-zero completeness, one loop clock, motion budget, exact identity assets, and approved typography when present.

render-qa

Produces deterministic static/GIF evidence without editing the artboard. It checks frame 0, feed-scale legibility, contrast, changed-pixel motion, seam, safe zone, duration, and file size. It returns build/render-report.json.

post-critic

Performs adversarial review before independent verification. It explicitly checks anti-slop, design taste, hooked-design-copy, creative-payoff, clean-creative-structure, verified-identity-assets, intentional-typography, restrained-palette, center-first-composition, UI fidelity, mascot identity, motion meaning, and evidence safety. It returns build/critic-report.json.

story-verifier

Read-only independent acceptance worker. It inspects artifact evidence directly, including identity provenance and typography loading, records one evidence row per criterion, enforces evidence-traceability and bounded-verification, and returns build/verification-report.json with PASS, FAIL:fixable, or FAIL:escalate.

Focused publication agent

masterone

Front-door onboarding and routing specialist. It diagnoses workspace profile state, identifies missing persistent preferences, and returns a routing-ready profile assessment to the parent workflow.

community-publisher

Runs only after the share-demo parent workflow has explicit consent, final verification PASS, rights confirmation, and a validated three-file export. It handles the contributor fork, fresh community/<user>/<slug> branch, scoped commit, push, and pull request against upstream main. It never merges, enables auto-merge, pushes to upstream main, or claims success without a real PR URL. Every contribution remains pending maintainer manual review and merge.

Coordination contract

The canonical complete sequence is:

design-study -> evidence-checker -> asset-curator -> creative-director -> story-architect -> palette-curator -> type-curator -> copy-compressor -> layout-composer -> caption-writer -> artboard-builder -> motion-director -> optional mascot-animator -> motion-engineer -> render-qa -> post-critic -> story-verifier

The sequence is declared in architecture/plugin-graph.json and mirrored by the create-post route. The community-publisher is intentionally outside that critical sequence and is reachable only through the optional share-demo route. Parent workflows own sequencing, user approvals, HOLD resolution, and bounded repair/export rules.

Validate all worker contracts with:

python3 scripts/plugin_graph.py check
python3 scripts/ecosystem_doctor.py check
python3 -m unittest tests.test_agent_contracts -v

MasterOne

masterone is the onboarding/profile worker used by the front-door masterone skill. It normalizes persistent preferences, reports missing blocking fields, separates discovered asset candidates from confirmed user choices, and recommends a canonical route back to the parent workflow. It does not orchestrate peer agents.

Antigravity Subagents Catalog

For Google Antigravity and Gemini CLI, all 19 canonical agents are declared in .agents/agents/catalog.json and managed via scripts/antigravity_agents.py. Each agent definition configures explicit write permissions, model tiers, and detailed system instructions matching the canonical contracts. See antigravity.md for full details.

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.