mindroom: Skill for Claude Code

Review documentation for accuracy, completeness, relevance, and consistency. Focus on things that require judgment—automated checks handle the rest. The Documentation Policy in AGENTS.md defines who reads docs/ and what belongs there.

What's Already Automated

Don't waste time on these—CI and pre-commit hooks handle them:

• README CLI output: markdown-code-runner regenerates CLI help blocks via docs/run_markdown_code_runner.py
• Linting/formatting: Handled by pre-commit

What This Review Is For

Focus on things that require judgment:

1. Accuracy: Does the documentation match what the code actually does?
2. Completeness: Are there undocumented features, config options, user-visible behaviors, or operator procedures? In user docs, internal limits, hardening, and failure or recovery paths are gaps only when a user would plausibly ask about them; docs/architecture/ pages keep the inventories AGENTS.md requires complete.
3. Relevance: Does every sentence in user docs pass the Documentation Policy, or does it describe implementation mechanics or change history? docs/architecture/ pages and the design records in docs/dev/ hold only what the code cannot show, as the AGENTS.md Documentation Policy lists; outside the inventories AGENTS.md requires, a sentence there that describes what the code does is a finding.
4. Clarity: Would a MindRoom agent answering a user question find and apply this? Are examples realistic?
5. Consistency: Do different docs contradict each other or repeat the same fact?
6. Freshness: Has the code changed in ways the docs don't reflect?

Review Process

1. Check Recent Changes

# What changed recently that might need doc updates?
git log --oneline -20 | grep -iE "feat|fix|add|remove|change|option"

# What code files changed?
git diff --name-only HEAD~20 | grep "\.py$"

Look for new features, changed defaults, renamed options, or removed functionality.

2. Verify Configuration Docs Against Code