agentleFS
imMamdouhaboammarimMamdouhaboammar/get-fable/skills/fable-domain/SKILL.md

fable-domain

Skill6 starsChanged 3 months agoMITView on GitHub

Discuss a domain with the user, research it from real sources, then generate a trusted skill bundle for it - a step-by-step workflow with a flowchart, a domain adapter, a trap fixture, and a smoke eval. Use when the user says "/fable-domain <sector>", "make a skill for <domain>", "add a domain to the fable method", or "give a lesser model Fable's workflow for <domain>". The bundle is the deliverable; a workflow without its flowchart, sources, and trap is not done.

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

Show its rules and text

We split this skill into the rules it gives an agent

15 rules. About 0 are common, 10 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈10

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

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

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

    When the user asks to add or generate a domain adapter (/fable-domain <sector>).

    line 43
  2. Distinctive:

    When expanding Fable's methodology to non-coding professional sectors.

    line 44
  3. Distinctive:

    When creating a domain-specific trap fixture and evaluation suite.

    line 45
  4. Distinctive:

    When the requested domain requires professional human licensure or presents high-harm risks (medical diagnosis, formal legal advice, …

    line 49
  5. Very specific:

    When the domain does not differ fundamentally from standard software engineering (use fable-execute or fable-plan).

    line 50
  6. Very specific:

    Read all existing adapters in references/domains/ and the governing template.

    line 70
  7. Very specific:

    Write the adapter according to references/domains/TEMPLATE.md.

    line 73
  8. Distinctive:

    Run repo checks to ensure structural integrity.

    line 77
  9. Distinctive:

    Run smoke eval comparing bare model behavior vs adapter-directed behavior.

    line 78
  10. Very specific:

    Run fable-judge adversarial audit over the bundle's claims and sources.

    line 79
  11. Distinctive:

    Use web search to fetch authoritative primary documents and regulatory texts.

    line 90
  12. Distinctive:

    Do not invent fictional regulatory citations or synthetic policy numbers.

    line 91
  13. Distinctive:

    If web access is unavailable to verify real-world regulations, stop generation rather than hallucinating standards.

    line 102
  14. Distinctive:

    If the smoke eval indicates no behavioral difference on the trap fixture, report the adapter as unproven.

    line 103
  15. Distinctive:

    All four bundle deliverables created and verified: workflow flowchart, adapter, trap fixture, and smoke eval.

    line 107

≈ “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.

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

Fable Domain

The fable-method ships domain adapters that translate its loop into a sector's nouns. This skill creates a new adapter and delivers a usable, step-by-step workflow with a flowchart for the domain.

Purpose

Generate validated, source-backed domain adapters and trap fixtures for new professional sectors (finance, marketing, legal compliance, devops, etc.). Translate general agent lifecycle discipline into domain-specific evidence standards and failure fraud tables.

When to Use

  • When the user asks to add or generate a domain adapter (/fable-domain <sector>).
  • When expanding Fable's methodology to non-coding professional sectors.
  • When creating a domain-specific trap fixture and evaluation suite.

When NOT to Use

  • When the requested domain requires professional human licensure or presents high-harm risks (medical diagnosis, formal legal advice, investment advice) — refuse and route to a qualified human.
  • When the domain does not differ fundamentally from standard software engineering (use fable-execute or fable-plan).
  • Generic skill authoring without domain-specific adapters (use fable-skill-creator).

Inputs

  • sector_name: The industry or functional discipline requested (e.g. finance, marketing, data-analysis).
  • user_requirements: Real-world use cases, trusted authorities, and non-negotiable boundaries provided by the user.

Expected Outputs

  1. Domain Workflow with Flowchart: Step-by-step approach distilled from research, plus a mermaid flowchart.
  2. Domain Adapter: Conforming to references/domains/TEMPLATE.md with cited primary sources.
  3. Trap Fixture: Evaluation scenario in eval/scenarios/ with GROUND-TRUTH.md defining the sector's central fraud.
  4. Smoke Eval: Initial control-vs-adapter validation runs.

Procedure

  1. Stage 1: Discuss: Adaptively clarify the use case, practitioner standards, trusted authorities, and forbidden actions. Enforce Red-lines (refusal for medical/legal/financial advice) and Scope stops (refusal if domain is standard software engineering).
  2. Stage 2: Research: Perform bounded web research against authoritative sources. Extract regulations, platform policies, practitioner evidence standards, and documented failure modes.
  3. Stage 3: Generate the Bundle:
    • Read all existing adapters in references/domains/ and the governing template.
    • Define exact sector scope boundaries.
    • Author the ordered workflow and mermaid flowchart.
    • Write the adapter according to references/domains/TEMPLATE.md.
    • Wire all routing surfaces (registry, catalog, docs).
    • Build the trap fixture with objective violation markers.
  4. Stage 4: Verify & Smoke-Eval:
    • Run repo checks to ensure structural integrity.
    • Run smoke eval comparing bare model behavior vs adapter-directed behavior.
    • Run fable-judge adversarial audit over the bundle's claims and sources.
    • Deliver outcome-first completion report.

Decision Rules

  • Red-Line Refusal: If a domain involves clinical medicine, formal legal counsel, or financial trading advice, refuse generation immediately.
  • Scope Gate: If a requested sector's evidence consists merely of files and stack traces, stop and declare that core coding skills already cover it.
  • Primary Source Mandate: Every regulatory threshold, compliance statute, or industry rule must cite a verified web source with access date.

Tool Policy

  • Use web search to fetch authoritative primary documents and regulatory texts.
  • Do not invent fictional regulatory citations or synthetic policy numbers.
  • Generate standard Mermaid diagrams for all workflow visualizations.

Evidence Requirements

  • Primary source URLs and verification dates in the adapter's Sources section.
  • Verifiable trap fixture with reproducible failing/passing test runs.
  • Passing fable-judge evaluation report.

Failure Handling

  • If web access is unavailable to verify real-world regulations, stop generation rather than hallucinating standards.
  • If the smoke eval indicates no behavioral difference on the trap fixture, report the adapter as unproven.

Completion Criteria

  • All four bundle deliverables created and verified: workflow flowchart, adapter, trap fixture, and smoke eval.
  • Zero red-line safety violations.
  • Registry and documentation references cleanly synchronized.

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.