Agent Skills: CLAUDE.md Manager

>-

UncategorizedID: oakensoul/aida-core-plugin/claude-md-manager

Install this agent skill to your local

pnpm dlx add-skill https://github.com/oakensoul/aida-core-plugin/tree/HEAD/skills/claude-md-manager

Skill Files

Browse the full folder contents for claude-md-manager.

Download Skill

Loading file tree…

skills/claude-md-manager/SKILL.md

Skill Metadata

Name
claude-md-manager
Description
>-

CLAUDE.md Manager

Manages CLAUDE.md configuration files that provide project-specific guidance to Claude Code. Supports the full CLAUDE.md hierarchy: project, user, and plugin scopes.

Activation

This skill activates when:

  • User invokes /aida claude create
  • User invokes /aida claude optimize
  • User invokes /aida claude validate
  • User invokes /aida claude list
  • CLAUDE.md management is needed

Operations

Create

Generate a new CLAUDE.md file at the specified scope. Detects project context (languages, tools, commands) and renders from templates.

Optimize

Audit an existing CLAUDE.md against best practices, generate findings with fix suggestions, and optionally apply fixes. Reports a quality score (0-100).

Validate

Run validation checks on existing CLAUDE.md files: structure, consistency, best practices, and alignment with detected project facts. Returns results without modifications.

List

Find all CLAUDE.md files in the hierarchy and report their validation status.

Two-Phase API

All operations follow the two-phase pattern:

Phase 1: Get Questions

Analyze context and return questions that need user input.

python {base_directory}/scripts/manage.py --get-questions \
  --context='{"operation": "create", "scope": "project"}'

Returns:

{
  "questions": [],
  "inferred": {
    "name": "my-project",
    "languages": ["Python"],
    "tools": ["Git", "pytest"],
    "commands": [{"command": "make test", "description": "Run tests"}]
  },
  "existing": null
}

Phase 2: Execute

Execute the operation with provided context and responses.

python {base_directory}/scripts/manage.py --execute \
  --context='{"operation": "create", "scope": "project"}' \
  --responses='{"name": "my-project"}'

Returns:

{
  "success": true,
  "message": "Created CLAUDE.md at ./CLAUDE.md",
  "path": "./CLAUDE.md"
}

CLAUDE.md Scopes

| Scope | Path | Use Case | | --- | --- | --- | | project | ./CLAUDE.md | Project documentation | | user | ~/.claude/CLAUDE.md | Global user preferences | | plugin | .claude-plugin/CLAUDE.md | Plugin documentation |

Path Resolution

Base Directory: Provided when skill loads via <command-message> tags.

Script Execution:

{base_directory}/scripts/manage.py

Resources

Scripts

| Script | Purpose | | --- | --- | | scripts/manage.py | Two-phase API entry point | | scripts/operations/claude_md.py | Core CLAUDE.md operations | | scripts/operations/utils.py | Shared utility re-exports |

References

| Document | Purpose | | --- | --- | | references/claude-md-workflow.md | End-to-end workflow guide | | references/best-practices.md | Best practices and scoring |

Templates

| Template | Purpose | | --- | --- | | templates/claude-md/project.md.jinja2 | Project-scope template | | templates/claude-md/user.md.jinja2 | User-scope template | | templates/claude-md/plugin.md.jinja2 | Plugin-scope template |