skills/doc-maintenance/SKILL.md
Systematic documentation audit and maintenance. This skill should be used when documentation may be stale, missing, or misorganized — after feature work, refactors, dependency upgrades, or as a periodic health check. It prescribes folder structure for docs/ and manual/, dispatches haiku subagents for codebase/doc scanning, and routes doc creation to specialized agents (reference-builder, technical-writer, learning-guide) with docs-architect as quality gate.
npx skillsauth add nickcrew/claude-cortex doc-maintenanceInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
Systematically audit, organize, and remediate project documentation by comparing the codebase against existing docs to find staleness, gaps, and misorganization.
scripts/doc_audit.py is run manually and reports findingsPhase 1: Audit → Run deterministic scan + haiku search agents
Phase 2: Triage → Classify findings by severity and action type
Phase 3: Remediate → Dispatch specialized agents to fix/create docs
Phase 4: Quality → docs-architect reviews all changes
Execute the bundled audit script to get a baseline report:
python3 skills/doc-maintenance/scripts/doc_audit.py
The script produces a structured report covering:
[text](path) pointing to missing files)docs/ or manual/)Pass --json for machine-readable output. Pass --root PATH to override project root detection.
After the deterministic scan, launch parallel haiku subagents to perform deeper analysis.
Use model: "haiku" on all Task calls in this phase to minimize cost.
See references/agent-dispatch.md for full prompt templates.
Agent 1 — Code-to-doc coverage scan (subagent_type: "Explore", model: "haiku"):
Search the codebase for public APIs, CLI commands, config schemas, and exported modules.
Cross-reference against existing docs. Report anything undocumented.
Agent 2 — Doc-to-code freshness scan (subagent_type: "Explore", model: "haiku"):
Read each doc file and verify the code constructs it references still exist and match
current signatures/behavior. Report mismatches.
Agent 3 — Structure compliance scan (subagent_type: "Explore", model: "haiku"):
Compare current docs/ and manual/ layout against the prescribed folder structure
in references/folder-structure.md. Report missing folders, misplaced files, naming violations.
Agent 4 — Diagram opportunity scan (subagent_type: "Explore", model: "haiku"):
Scan all markdown files for ASCII/text diagrams (box-drawing characters, arrow notation,
indented tree structures beyond a few simple nodes) that should be converted to Mermaid.
Also identify sections describing flows, architectures, state machines, sequences, or
relationships where a diagram would add clarity but none exists. Report the file path,
line range, diagram type (flowchart, sequence, state, ER, etc.), and whether it is a
conversion or a net-new diagram.
Launch all four agents in parallel.
Combine the script output with agent findings into a single audit report. Deduplicate overlapping findings. The report becomes the input for Phase 2.
Classify each finding into one of these action categories:
| Category | Description | Example |
|----------|-------------|---------|
| stale | Doc exists but references outdated code/behavior | CLI flag renamed but docs show old name |
| missing | No doc exists for a documented-worthy item | Public API endpoint with no reference doc |
| orphan | Doc exists but is unreachable / unlinked | Guide file not in any index or nav |
| misplaced | Doc exists but is in the wrong folder | Tutorial sitting in docs/architecture/ |
| irrelevant | Doc covers removed functionality | Guide for a deleted feature |
| structural | Folder structure deviates from prescribed layout | Missing docs/security/ folder |
| diagram-convert | ASCII/text diagram should be Mermaid | Complex box-drawing flowchart in architecture doc |
| diagram-missing | Section would benefit from a diagram | Multi-step process described only in prose |
Assign severity:
Route each finding to the appropriate specialist agent. Use the Task tool with
the subagent types listed below. See references/agent-dispatch.md for detailed
prompt templates.
| Doc type | Subagent type | Target location |
|----------|---------------|-----------------|
| API reference docs | reference-builder | docs/reference/ or docs/api/ |
| Architecture docs | technical-writer | docs/architecture/ |
| Developer guides (style, local dev, workflows) | technical-writer | docs/development/ |
| Testing docs | technical-writer | docs/testing/ |
| Security docs | technical-writer | docs/security/ |
| User-facing tutorials | learning-guide | manual/tutorials/ |
| User-facing how-to guides | learning-guide | manual/guides/ |
| User-facing getting started | learning-guide | manual/getting-started/ |
| Plans and proposals | technical-writer | docs/plans/ |
| ASCII diagram conversion | mermaid-expert | Inline in existing doc |
| New diagrams for prose sections | mermaid-expert | Inline in existing doc |
Parallel dispatch: Group independent remediation tasks and dispatch them simultaneously. Only serialize when one doc depends on another (e.g., an API reference needed before a tutorial that links to it). Dispatch up to 4 remediation agents in parallel per batch.
For updates to existing docs: Provide the agent with the current file contents and the specific finding to fix. Instruct it to make minimal, targeted edits.
For new docs: Provide the agent with the relevant source code, the target file path, and the folder-structure spec so it follows naming conventions.
After all remediation agents complete, dispatch a single docs-architect agent to review
the full set of changes. The quality gate checks:
If the quality gate fails, loop back to Phase 3 for the specific issues flagged. Maximum 2 remediation loops before escalating to the user.
The prescribed folder layout is defined in references/folder-structure.md. Summary:
docs/ — Internal / developer documentationdocs/
├── architecture/ — System design, ADRs, component diagrams
├── development/ — Developer guides: style, local setup, issue tracking
├── plans/ — Proposals, RFCs, roadmaps
├── reviews/ — Code review records, audit reports
├── testing/ — Test strategy, coverage reports, test plans
├── reports/ — Generated reports, metrics, analysis
├── security/ — Security policies, threat models, audit findings
├── api/ — Internal API docs (OpenAPI specs, gRPC protos)
├── reference/ — CLI reference, config reference, manpages
├── ideas/ — Exploratory notes, spikes, brainstorms
└── archive/ — Deprecated docs preserved for history
manual/ — User-facing documentation (project root)manual/
├── getting-started/ — Installation, quickstart, first steps
├── guides/ — How-to guides for common tasks
├── tutorials/ — Step-by-step learning paths
├── reference/ — User-facing command/config reference
└── troubleshooting/ — FAQ, common errors, known issues
README.md — Project rootThe main README is audited for accuracy but not reorganized. Findings about the README are reported as stale/missing items for manual remediation.
docs/ and manual/ — link insteaddocs/ or developer docs into manual/development
Product vision, roadmap development, and go-to-market execution with structured prioritization frameworks. Use when evaluating features, planning product direction, or assessing market fit.
development
Complete operational workflow for implementer agents (Codex, Gemini, etc.) making code changes and writing tests. Drives all work through atomic commits — each loop operates on the smallest complete, reviewable change. Defines the Code Change Loop, Test Writing Loop, Lint Gate, and Issue Filing process with circuit breakers, severity levels, and escalation rules. Requires `cortex git commit` for all commits. Includes bundled provider-aware review scripts that keep same-model shell-outs as the last resort, plus a fresh-context Codex fallback for code review and test audit. Use this skill when starting any implementation task.
development
Use this skill when writing product requirements documents, prioritizing features, creating user stories, defining acceptance criteria, or setting product metrics. Trigger phrases: 'write a PRD for', 'prioritize this feature backlog', 'write user stories for', 'help me define acceptance criteria', 'what metrics should we track for'. Not for writing code, designing UI mockups, or conducting user research interviews.
tools
Automates browser interactions for web testing, form filling, screenshots, and data extraction. Use when the user needs to navigate websites, interact with web pages, fill forms, take screenshots, test web applications, or extract information from web pages.