SuperAPI AI Directory

discover-docs-norms

Writing & research

View repository

Updated 10-16 06:50

About this skill

Skill: Discover Docs Norms

Purpose

Discover and propose project-level documentation norms (paths, naming, front-matter, lifecycle), providing input basis for subsequently creating or updating norm files.


Core Objective

Primary objective: produce a reviewable norm proposal, without executing norm persistence.

Success criteria(all must be satisfied):

  1. ✅ Scanned the target docs scope and extracted existing conventions
  2. ✅ Output a norm proposal (including path mapping, naming rules, front-matter rules, confidence)
  3. ✅ Clearly identified conflicts and low-confidence items
  4. ✅ Provided next-step handoff recommendations (define-docs-norms)
  5. ✅ Did not write or overwrite docs/ARTIFACT_NORMS.md

Acceptance test: Can the team review and decide whether to persist based solely on the proposal, without modifying the repository structure?


Scope Boundaries

This skill is responsible for:

  • Scanning the docs structure and inferring documentation governance conventions
  • Forming proposals and marking confidence and conflicts
  • Outputting suggested canonical rules and migration hints

This skill is not responsible for:

  • Creating or updating docs/ARTIFACT_NORMS.md(use define-docs-norms)
  • Batch organizing file structure (use tidy-repo)
  • Compliance assessment and health scoring (use assess-docs)

Handoff point: after the proposal is confirmed, hand off to define-docs-norms to persist the norms.


Use Cases

  • Before the project's first governance effort, existing documentation conventions need to be identified first
  • When norms are disputed, a discussable candidate solution needs to be generated
  • In multi-team collaboration, unify candidate norms first before deciding whether to write them

Behavior

Phase 1: Scanning and Sampling

  1. Scan docs/(if it does not exist, scan Markdown in the repository)
  2. Read front-matter, paths, file names, titles
  3. Extract artifact_type, naming patterns, date patterns, directory distribution

Phase 2: Rule Derivation

  1. Generate candidate artifact_type -> path_pattern mappings
  2. Derive naming rules (kebab/snake/camel, date prefix strategy)
  3. Derive the minimal front-matter field set
  4. Identify conflict distribution (multiple paths for the same type, multiple naming styles)

Phase 3: Confidence and Risk Assessment

  1. Score each rule (0-100)
  2. Mark low-confidence (< 70) rules
  3. Provide conflict resolution recommendations (which path to keep, which files to migrate)

Phase 4: Output Proposal (read-only)

  1. Generate docs/calibration/docs-norms-proposal.md
  2. Output proposal structure: rule table, conflict table, migration recommendations, decision question list
  3. Clarify next step: if it needs to take effect, execute define-docs-norms

Input & Output

Input

  • Project path (defaults to the current repository)
  • Optional docs root path
  • Optional confidence threshold

Output

  • docs/calibration/docs-norms-proposal.md(norm proposal)
  • Does not create/update docs/ARTIFACT_NORMS.md

Restrictions

Hard Boundaries

  • Without explicit user request, this skill must not write docs/ARTIFACT_NORMS.md
  • This skill must not perform repository organization or file moves
  • This skill must not disguise proposals as already-effective specifications

Skill Boundaries

Do not do these (other skills are responsible):

  • Specification persistence and changes:define-docs-norms
  • Structural organization:tidy-repo
  • Compliance and readiness assessment:assess-docs

Self-Check

  • Scanned the target docs scope and extracted rule candidates
  • Output path/naming/front-matter proposals
  • Marked low-confidence and conflicting items
  • Provided handoff recommendations define-docs-norms
  • Not written docs/ARTIFACT_NORMS.md

Examples

Example 1: Only proposals found

  • Input: Existing project docs/ structure
  • Output:docs/calibration/docs-norms-proposal.md
  • Follow-up: After review approval, call define-docs-norms

Example 2: Multiple conflicting directory sets exist

  • Input: Both docs/design/ and docs/design-decisions/
  • Output: Conflict adjudication recommendations + migration checklist
  • Follow-up: By define-docs-norms solidify canonical rules

Appendix: Output Contract

This skill produces a Docs Norms Proposal (does not write the specification itself):

Element Format Required fields Path pattern
Proposal body Markdown front-matter (artifact_type=proposal / lifecycle=snapshot); sections: scan results / proposed rules / differences from current state / risks docs/calibration/docs-norms-proposal.md
Proposed rules Table rule_kind(path / naming / lifecycle)/ pattern / evidence_files / confidence "Proposed rules" section
Differences from current state List items Instance paths in existing docs that violate this proposal; used to assess change cost "Differences from current state" section