up-repo

/up-docs:repo

Update the active repository's documentation files to reflect work done in the current session.

Workflow

1. Assess Session Context

Gather what changed during the session:

git diff --stat HEAD~5 HEAD 2>/dev/null || git diff --stat
git log --oneline -10

Also consider the conversation history: files discussed, features implemented, bugs fixed, configuration changes made. Build a mental model of what the session accomplished.

2. Locate Documentation Files

Read the project CLAUDE.md for any ## Documentation section that specifies which files to update.

If no explicit mapping exists, discover documentation files:

# Find all markdown docs in standard locations
find . -maxdepth 1 -name "*.md" -type f
find ./docs -name "*.md" -type f 2>/dev/null

Common targets: README.md, CLAUDE.md, CHANGELOG.md, docs/*.md, and any other .md files at the project root or in docs/.

3. Read and Evaluate Each File

Read every candidate documentation file. For each one, determine whether the session's changes make any part of it stale or incomplete.

Evaluation criteria:

• Does the file describe features, commands, or configuration that changed this session?
• Are there sections that reference files, functions, or behavior that no longer match?
• Is anything new from this session missing from the documentation?

Skip files that are clearly unaffected by the session's work.

4. Draft and Apply Updates

For each file that needs changes:

• Read the file's current content in full
• Preserve the existing tone, structure, and formatting
• Make targeted edits; do not rewrite sections that are still accurate
• Add new content where the session introduced something not yet documented
• Remove or correct content that the session's changes have invalidated

Do not add boilerplate, badges, or sections that the file doesn't already have. Match the document's existing conventions.

5. Summary Report

Read ${CLAUDE_PLUGIN_ROOT}/templates/summary-report.md for the output format.

Emit the summary report using the Repo layer format. Every file examined gets a row in the table, including files where no changes were needed.