claude-md-improver

name: claude-md-improver description: >- Audit and improve Claude Code CLAUDE.md guidance files, .claude/rules/ path-scoped rules, and companion code_map.md navigation maps. Use when the user asks to check, audit, update, optimize, or fix CLAUDE.md files; mentions nested CLAUDE.md, additive loading, ancestor/descendant loading, path-scoped rules, @import chains, CLAUDE.local.md, claudeMdExcludes, or code_map.md; or says "优化 CLAUDE.md", "审计 CLAUDE.md", "检查嵌套 CLAUDE.md", "Claude Code 项目指导", "生成 code_map (Claude)". Make sure to use this skill whenever the user wants to manage Claude Code memory files at any layer, even if they only mention CLAUDE.md without explicit audit wording. version: 1.0.0 category: developer-tools-integrations tags: - claude-code - claude-md - repository-guidance - memory - audit - documentation - code-map argument-hint: "[audit-or-update-goal]" allowed-tools: Read, Glob, Grep, Edit, Write, Bash(git *), Bash(find *), Bash(Get-ChildItem *), Bash(Get-Content *), Bash(Select-String *)

CLAUDE.md Improver

Audit and improve Claude Code CLAUDE.md guidance files, .claude/rules/ path-scoped rules, and the companion code_map.md navigation map so future Claude Code sessions receive concise, accurate, additively-layered project instructions and fast code search entry points.

Default mode is report-first. Output a quality report and proposed diff before writing. If the user explicitly asks to implement an approved plan, continue directly to targeted edits and verification.

Core Semantics

Claude Code memory files use additive layering, not directory-scoped override. This is the most common point of confusion when migrating instructions from Codex AGENTS.md or other agents.

• Claude Code walks up the directory tree from the current working directory; every CLAUDE.md and CLAUDE.local.md found along the way is loaded at session start and concatenated into context, ordered from filesystem root to working directory.
• CLAUDE.md files in descendant directories are loaded lazily when Claude reads files in those directories during the session.
• CLAUDE.md files in sibling directories are never loaded for the current session.
• Files load additively: all layers contribute to context simultaneously. There is no "deeper file overrides parent" semantic; more specific guidance simply sits later in the prompt.
• Precedence chain at session start: managed policy → user (~/.claude/CLAUDE.md) → project (./CLAUDE.md then ./.claude/CLAUDE.md) → CLAUDE.local.md per layer. Repository CLAUDE.md is project guidance. User-level ~/.claude/CLAUDE.md is personal preference and is not edited unless the user explicitly asks.
• @path imports expand at launch (recursive, max depth 5). They do not save context; they only organize files.
• Block-level HTML comments (<!-- ... -->) are stripped before injection. Use them for human-maintainer notes that should not consume tokens.
• .claude/rules/*.md with a paths: frontmatter glob loads only when Claude reads files matching that glob. Rules without paths: load unconditionally.
• CLAUDE.md carries durable behavioral constraints, commands, and project conventions. code_map.md carries navigational structure, search anchors, entry points, and generated/ignored directory notes.
• Every created or updated CLAUDE.md should name the relative code_map.md path agents must read before broad grep, for example Before broad grep, read ./code_map.md.
• The Claude Code official guidance and the large-codebases best-practices post both recommend keeping each CLAUDE.md under 200 lines; longer files reduce adherence and inflate every session's startup tokens.
• Preserve hook-managed marker blocks such as <!-- OMX:RUNTIME:START --> ... <!-- OMX:RUNTIME:END --> and <!-- OMX:TEAM:WORKER:START --> ... <!-- OMX:TEAM:WORKER:END -->. These are stripped at injection but are still meaningful to hook tooling and must survive edits intact.

See references/claude-md-loading.md for the full loading model, including edge cases around --add-dir, claudeMdExcludes, and auto memory boundaries.

Workflow

Phase 1: Discovery

Find every memory file in scope.

# POSIX shells
find . \( -name CLAUDE.md -o -name CLAUDE.local.md -o -name code_map.md \) \
  -not -path './.git/*' \
  -not -path './node_modules/*' \
  -not -path './target/*' \
  -not -path './dist/*' \
  -not -path './build/*' \
  -not -path './.omx/state/*'
find . -path '*/.claude/rules/*.md' -not -path './node_modules/*'

# PowerShell
Get-ChildItem -Recurse -Force -Include CLAUDE.md,CLAUDE.local.md,code_map.md |
  Where-Object { $_.FullName -notmatch '\\.git|node_modules|target|dist|build|\\.omx\\state' } |
  Select-Object -ExpandProperty FullName
Get-ChildItem -Recurse -Force -Path .\.claude\rules\*.md -ErrorAction SilentlyContinue

Exclude generated, vendored, dependency, cache, and build-output directories from creation scans: .git/, node_modules/, target/, dist/, build/, .omx/state/, vendor/, coverage output, generated docs/site output, and language-specific package caches.

Also note, but do not edit by default:

~/.claude/CLAUDE.md
~/.claude/rules/
/Library/Application Support/ClaudeCode/CLAUDE.md   # macOS managed policy
/etc/claude-code/CLAUDE.md                          # Linux/WSL managed policy
C:\Program Files\ClaudeCode\CLAUDE.md               # Windows managed policy

Discover candidate subtrees for new nested CLAUDE.md (and possibly local code_map.md) before proposing writes. Score only real source subtrees that show one or more of: distinct stack or framework, independent manifest/command surface, divergent conventions, frequent agent work in that area, local safety boundary, or pressure to relieve a bloated root CLAUDE.md.

Classify each file:

Phase 2: Evidence Collection

For every in-scope file, verify claims against the repository:

• commands referenced in package.json, justfile, Cargo.toml, pyproject.toml, Makefile, CI workflows
• actual directory structure and entry points named in the file
• @import targets: every @path/to/file must resolve relative to the file containing it; recursion depth ≤ 5
• code_map.md existence and whether each CLAUDE.md map pointer uses an explicit relative path
• .claude/rules/*.md frontmatter: paths: globs are valid, file globs match real source files, conditional rules are not duplicating root content
• alignment with .claude/settings.json — do not duplicate hard permissions.deny constraints as soft prose; if the setting already blocks an action, do not also rephrase it as a rule
• claudeMdExcludes settings (in .claude/settings.json, .claude/settings.local.json, or managed) and whether they conflict with the files you are auditing
• bloat signals: total line count vs the 200-line target, repeated content across nested files, code restatement that belongs in code_map.md
• preserved hook marker blocks; flag any unbalanced OMX:...:START / END pairs
• CLAUDE.local.md presence — confirm it is listed in .gitignore; do not read its content

Phase 3: Quality Assessment

Use references/quality-criteria.md for detailed scoring, including the nested CLAUDE.md creation scorecard.

Quick checklist (100 points total):

Grades:

• A (90-100): layered, current, executable, concise
• B (70-89): useful with minor gaps
• C (50-69): basic but missing key operational detail
• D (30-49): sparse, stale, or confusing
• F (0-29): missing, misleading, or counterproductive

For each candidate nested subtree, record creation score and decision: create (≥60), candidate only (40-59), or do not create (<40). For each existing nested file, note whether it could move into a .claude/rules/*.md with a paths: glob instead.

Phase 4: Report Before Editing

Always provide this report before edits unless the user already approved an implementation plan.

## CLAUDE.md Quality Report

### Summary
- Files found: X (CLAUDE.md: X, CLAUDE.local.md: X, .claude/rules: X, code_map: X)
- Project root guidance: present at <path>/missing
- Root code map: present/missing/shared with agents-md-improver
- Nested project files: X (X over 200 lines)
- Average score: X/100
- Files needing update: X
- Candidate nested guidance dirs: X create / X candidate-only / X skipped

### Layering Map
| File | Loads when | Adds | Notes |
|---|---|---|---|
| `CLAUDE.md` | session start (ancestor) | repo-wide pointers + gotchas | 142 lines |
| `packages/api/CLAUDE.md` | reading files under `packages/api/` | API stack, local commands, secrets boundary | 88 lines |
| `.claude/rules/testing.md` | reading `**/*.test.ts` | TDD rules | paths-scoped |

### Code Map Coverage
| Map | Covers | Referenced by | Notes |
|---|---|---|---|
| `code_map.md` | repo root | `CLAUDE.md`, `AGENTS.md` | shared with agents-md-improver |

### Nested Guidance Candidates
| Directory | Score | Decision | Evidence |
|---|---:|---|---|
| `packages/api/` | 75 | create `packages/api/CLAUDE.md` | distinct Python stack, own pytest commands, secrets boundary |

### File-by-File Assessment

#### 1. `CLAUDE.md`
**Score: XX/100 (Grade: X)**

| Criterion | Score | Notes |
|---|---:|---|
| additive layering clarity | X/20 | ... |
| executable commands and gates | X/20 | ... |
| architecture and routing | X/15 | ... |
| tool permissions and safety | X/15 | ... |
| Claude Code workflow fit | X/15 | ... |
| conciseness and currency | X/15 | line count XXX vs 200 target |

**Issues**
- ...

**Proposed changes**
- ...

Phase 5: Targeted Updates

When approved or already authorized by a plan, follow references/update-guidelines.md. The nine load-bearing rules:

1. Root CLAUDE.md holds repo-wide pointers and critical gotchas only; target under 200 lines.
2. Nested CLAUDE.md adds local stack, commands, and safety; never restates the parent content because parent content already loaded.
3. Behavior and constraints live in CLAUDE.md; navigation, anchors, and directory indexes live in code_map.md.
4. Path-specific rules belong in .claude/rules/*.md with paths: frontmatter — they load only when relevant and keep the root file lean.
5. Reference ./code_map.md with prose ("Before broad grep, read ./code_map.md") rather than @code_map.md. @import loads at launch and defeats the on-demand purpose of a map.
6. In the root CLAUDE.md, list known nested CLAUDE.md files under a "Nested Documentation" section so Claude can discover sibling subtrees by name even when their files are not yet loaded.
7. Preserve hook-managed marker blocks exactly. Preserve @AGENTS.md import bridges when the repository runs both Claude Code and Codex.
8. Remove generic LLM advice ("write clean code", "understand before changing"); these dilute the signal-to-noise ratio.
9. If CLAUDE.local.md is present, only confirm it is in .gitignore; do not read its content or copy it into shared files.

Phase 6: Verification

Run the smallest checks that prove the edits:

git diff --check

Manual checks to perform on every updated file:

• line count is under 200; otherwise propose splitting into .claude/rules/*.md
• every @path/to/file import target exists on disk; recursion depth ≤ 5
• every code_map.md pointer uses an explicit relative path
• every .claude/rules/*.md paths: glob is valid YAML and matches at least one real file
• no new content was added that duplicates .claude/settings.json hard rules
• hook marker blocks are intact with matched start/end pairs
• CLAUDE.local.md is in .gitignore if it exists

If a documented full gate is expensive (browser tests, deploys), state whether it was run or why it was skipped.

Reference Files

• references/claude-md-loading.md — Claude Code loading model, edge cases, and what each loading layer can and cannot do
• references/quality-criteria.md — scoring rubric, nested creation scorecard, and red flags
• references/templates.md — root, monorepo package, frontend, backend, docs, .claude/rules, @AGENTS.md bridge, and code_map.md templates
• references/update-guidelines.md — what to add, what to avoid, what to preserve, diff format

Common Issues to Flag

• root CLAUDE.md exceeds 200 lines because path-specific content was never moved to .claude/rules/*.md
• root CLAUDE.md missing an explicit ./code_map.md pointer when a root map exists or should exist
• nested CLAUDE.md restates parent content under the false assumption that nested files override the root
• nested CLAUDE.md only says "read the code map" without naming the relative map path
• CLAUDE.md bloated with directory index content that belongs in code_map.md
• low-score, generated, vendored, dependency, or build-output directories receiving unnecessary nested CLAUDE.md
• @import chains exceeding 5 hops or referencing missing files
• block-level <!-- ... --> comments used for instructions Claude is expected to follow (they are stripped before injection)
• .claude/rules/*.md without paths: frontmatter that would have been a perfect path-scoped fit
• CLAUDE.local.md checked into git (must be gitignored)
• Codex-only guidance from AGENTS.md pasted into CLAUDE.md instead of using the @AGENTS.md import bridge
• hook-managed marker blocks accidentally rewritten or with mismatched start/end pairs
• external production services or credentials not called out as a safety boundary
• skills, sub-agents, hooks, or MCP servers described as available when they are only aspirational

Final Output After Edits

## CLAUDE.md Update Summary

### Files changed
- `CLAUDE.md` — ...
- `packages/api/CLAUDE.md` — ...
- `.claude/rules/testing.md` — ...
- `code_map.md` — ...

### What improved
- layering clarity and root-vs-nested division of labor
- command/gate accuracy
- safety boundaries aligned with `.claude/settings.json`
- map-first search flow with explicit relative `code_map.md` paths
- root file shrunk from XXX to YYY lines

### Verification
- `git diff --check` — passed
- `<targeted command>` — passed/failed/skipped with reason

### Remaining risks
- ...