Skip to content
Back to skills

Mend Agent Rules

ASecurity

Synchronize CLAUDE.md and AGENTS.md with available agents and skills.

  • 57 stars
  • 0 votes
  • 0 copies
  • 0 views
  • Added October 17, 2026
documentationgodocumentation

Security analysis

A100/100

Scanned October 17, 2026

npx -y skills add diegosouzapw/awesome-omni-skill --skill mend-agent-rules --agent claude-code

Installs into .claude/skills of the current project.

Are you the author of Mend Agent Rules?

Add the live security badge to your README. It updates with every re-scan.

Security grade badge for Mend Agent Rules
[![Security: A — Skills Directory](https://www.skillsdirectory.com/api/skills/diegosouzapw-mend-agent-rules/badge)](https://www.skillsdirectory.com/skills/diegosouzapw-mend-agent-rules)

More formats (shields.io, HTML) on the badges page. Keep it an A: scan every change in CI with Pro.

Download with Pro
SKILL.md
---
name: mend-agent-rules
description: Synchronize CLAUDE.md and AGENTS.md with available agents and skills.
---

# Mend Agent Rules

This skill automates the synchronization of `CLAUDE.md` and `AGENTS.md` using AI / LLM to ensure they always reflect the current state of the repository's agents and skills.

## Purpose

To provide AI agents with up-to-date instructions and context about available specialized capabilities without manual maintenance.

## Instructions

1.  **Discovery**: Scan the following directories and categorize them:
    - **Project Skills**: All `SKILL.md` files in `.claude/skills/` (recursive).
    - **Agents**: Agent definitions or metadata in `.claude/agents/`.
2.  **Extraction**: For each skill/agent found, extract:
    - `name`: From YAML frontmatter or directory name.
    - `description`: From YAML frontmatter or first paragraph.
    - `path`: Relative path from the workspace root.
3.  **Target Identification**: Identify documentation files and their respective markers:
    - **CLAUDE.md** and **AGENTS.md**:
      - `<!-- AVAILABLE_SKILLS_START -->` and `<!-- AVAILABLE_SKILLS_END -->`
      - `<!-- AVAILABLE_AGENTS_START -->` and `<!-- AVAILABLE_AGENTS_END -->`
      - `<!-- FRAMEWORK_DOCS_START -->` and `<!-- FRAMEWORK_DOCS_END -->`
4.  **Formatting**:
    - **Skills List**: Generate a table of all discovered project skills.
    - **Agents List**: Generate a table of discovered agents.
5.  **Synchronization**: Replace the content between the specific markers in BOTH target files.
    - Update BOTH `CLAUDE.md` and `AGENTS.md` with the project skills table, agents table, and framework documentation.
    - **CRITICAL**: The framework documentation (explaining what Agent Skills are, directory structure, and guidelines) MUST be present in both files between the `FRAMEWORK_DOCS` markers. If it is missing from one but present in the other, synchronize it.
6.  **Verification**: Confirm that both files are identical except for their top-level headers (`# CLAUDE.md` vs `# AGENTS.md`) and that all markers are preserved.

## Markers

### CLAUDE.md

```html
<!-- AVAILABLE_SKILLS_START -->
<!-- AVAILABLE_SKILLS_END -->

<!-- AVAILABLE_AGENTS_START -->
<!-- AVAILABLE_AGENTS_END -->

<!-- FRAMEWORK_DOCS_START -->
<!-- FRAMEWORK_DOCS_END -->
```

### AGENTS.md

```html
<!-- AVAILABLE_SKILLS_START -->
<!-- AVAILABLE_SKILLS_END -->

<!-- AVAILABLE_AGENTS_START -->
<!-- AVAILABLE_AGENTS_END -->

<!-- FRAMEWORK_DOCS_START -->
<!-- FRAMEWORK_DOCS_END -->
```

## Example Table Format

<!-- markdown-link-check-disable -->

| Name     | Description                                                              | Link                                                                 |
| :------- | :----------------------------------------------------------------------- | :------------------------------------------------------------------- |
| lint-fix | Iteratively run linters, apply auto-fixes, and resolve remaining issues. | [.claude/skills/lint-fix/SKILL.md](.claude/skills/lint-fix/SKILL.md) |

<!-- markdown-link-check-disable -->

Attribution

Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.

Comments

Loading comments…

Newsletter

Get the weekly list of safe new skills, and the ones we pulled.