agentleFS
Cogni-AI-OUCogni-AI-OU/cogni-ai-agent-skills/fact-writer/SKILL.md

fact-writer

Skill4 starsChanged 6 months agoMITView on GitHub

Strict guidelines for creating, writing, and maintaining verifiable project fact files (e.g., FACTS.md or FACTS.mmd) without contradictions, ensuring state compression and lexical ordering. You MUST load this skill when managing canonical project fact files.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    When recording temporal, ephemeral state like "Build 45 failed" (fact stores are for persistent architectural truths).

    line 25
  2. Distinctive:

    Read existing fact files entirely before proposing updates.

    line 37

≈ “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
<!-- markdownlint-disable MD013 MD023 MD031 MD032 -->

Fact Writer

Guidance for structuring and maintaining verifiable project fact files. Use this skill when managing FACTS.md, FACTS.mmd, or similar canonical stores.

When to Use

  • Creating or updating canonical fact stores (e.g., FACTS.md, FACTS.mmd).
  • Documenting verifiable ecosystem properties, constraints, or architectural decisions.
  • Resolving state conflicts or contradictions within documentation limits.

When Not to Use

  • For writing conversational READMEs or unstructured tutorial documentation (use docs-writer instead).
  • When recording temporal, ephemeral state like "Build 45 failed" (fact stores are for persistent architectural truths).
  • For writing executable code or scripts.

Common Pitfalls

  • Violating Lexical Order: Appending a new fact at the bottom of the list instead of inserting it in strict alphabetical order, breaking the state compression logic.
  • Silent Contradictions: Adding a fact that explicitly contradicts an existing node without explicitly removing or updating the old node.
  • Prose Injection: Writing long paragraphs of explanation instead of dense, structured nodes.

Core Process

  1. Verify State: Read existing fact files entirely before proposing updates; never infer missing state.
  2. Contradiction Transparency: Surface conflicts immediately; reject silent overwrites when a proposed fact contradicts an existing A ≠ ¬A state. Replace, do not append.
  3. Verifiable Facts Only: Record objective facts grounded in repository configuration, code, and explicit documentation. Zero subjective prose.
  4. Structured Format: Enforce rigid hierarchical structures (e.g., Mermaid mindmap or perfectly nested Markdown lists). Group by logical domains like Architecture, Context, Ecosystem.
  5. Lexical Sorting: Insert new facts in strict alphabetical order within their domain/level.
    • Case: Case-insensitive (e.g., A and a share same priority; App before apple).
    • Numbers: Natural sort (e.g., v2 < v10).
    • Special Characters: Sort symbols (e.g., @, _) before alphanumeric characters.
  6. Reversibility Focus: Offload history, diffs, and rollback logic entirely to Git. Eradicate historical changelogs, sequence IDs, and obsolete states from the fact file itself.

Diagnostics and Revisions

  • Contradiction Sweep: Pre-commit self-review to guarantee zero logical conflicts.
  • Orphan Pruning: Seamlessly drop redundant, outdated, or superseded facts (e.g., deprecated libraries) to maintain high-density state compression.

What to Avoid

  • NEVER insert nodes or lists out of alphabetical order.
  • NEVER include conversational prose, narratives, or explanations in fact structures.
  • NEVER hallucinate project details. If unsupported by repo state, reject ingestion.
  • NEVER preserve historical states when superseded. Replace with the new fact.

Related Skills

  • mermaid: You MUST load this skill when creating or maintaining Mermaid.js diagrams.
  • mermaid-beta: You MUST load this skill when working with experimental Mermaid.js beta diagrams.

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.