agentleFS
JBurlisonJBurlison/WoWAddonAPIAgents/skills/validate-agent-files/SKILL.md

validate-agent-files

Skill75 starsChanged 13 days agoMITView on GitHub

Validates AI coding assistant customization files (agents, skills, prompts, instructions) for correct format and structure. Works with GitHub Copilot, Claude Code, Codex, OpenCode, and other providers. Use when checking if agent files are properly configured, troubleshooting agent issues, or before committing new customization files.

1 other project uses a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Before committing new agents, skills, prompts, or instructions

    line 16
  2. Distinctive:

    When an agent isn't behaving as expected

    line 17
  3. Very specific:

    After modifying any .github customization files

    line 19
  4. Distinctive:

    No usage examples in description

    line 70
  5. Distinctive:

    When validating all files, provide summary:

    line 181

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Validate Agent Files

Validates that agent, skill, prompt, and instruction files follow the correct format and structure.

Host-specific packaging

This is a repository maintenance skill, excluded from the WoW plugin. Maintain canonical WoW skills under skills and generate the package with npm run build. Agent metadata, rules, commands, and discovery paths differ by host; do not copy Copilot conventions into Codex or Claude mechanically. See file locations. Examples using <provider>/ below describe Copilot-style source files and must be adapted using host specifications.

When to Use

  • Before committing new agents, skills, prompts, or instructions
  • When an agent isn't behaving as expected
  • To audit existing customization files for issues
  • After modifying any .github customization files

Validation Process

Step 1: Identify File Type

Determine the type based on location and extension:

  • <provider>/agents/*.md → Agent file (user-invokable)
  • <provider>/agents/*.subagent.agent.md → Sub-agent file (workflow component)
  • <provider>/skills/*/SKILL.md → Skill file
  • <provider>/prompts/*.prompt.md → Prompt file
  • <provider>/instructions/*.instructions.md → Instruction file

Step 2: Apply Type-Specific Validation

Agent File Validation (<provider>/agents/*.md)

Required Structure:

---
name: agent-name
description: When to use this agent (should include examples)
user-invokable: true  # Optional, defaults to true
---

[System prompt body]

Supported Frontmatter Attributes:

  • name (required) - Agent identifier
  • description (required) - When/how to use, with examples
  • user-invokable (optional) - Set to false for sub-agents (default: true)
  • tools - List of allowed tools
  • model - Specific model to use
  • handoffs - Other agents this can delegate to

Checks:

  1. ✓ YAML frontmatter present with --- delimiters
  2. ✓ name field exists and is non-empty
  3. ✓ description field exists (recommend 50+ characters with examples)
  4. ✓ Body content exists after frontmatter
  5. ✓ If tools specified, they are valid tool names
  6. ✓ If filename contains .subagent.agent.md, verify user-invokable: false is set

Naming Convention Checks:

  • User-facing agents: <name>.agent.md or <name>.md
  • Sub-agents: <name>.subagent.agent.md with user-invokable: false

Common Issues:

  • Missing --- delimiters
  • Empty or minimal description
  • No usage examples in description
  • Body content missing or too brief
  • Sub-agent missing user-invokable: false
  • Sub-agent not using .subagent.agent.md naming convention

Skill File Validation (<provider>/skills/*/SKILL.md)

Required Structure:

---
name: skill-name
description: What this skill does and when to use it.
---

[Skill instructions body]

Supported Frontmatter Attributes:

  • name (required) - Must match parent directory name, lowercase with hyphens
  • description (required) - Max 1024 chars, describes function and triggers
  • license (optional) - License information
  • compatibility (optional) - Environment requirements
  • metadata (optional) - Key-value pairs for additional info
  • allowed-tools (optional) - Space-delimited pre-approved tools

Checks:

  1. ✓ File is named SKILL.md inside a directory
  2. ✓ name matches parent directory name exactly
  3. ✓ name is lowercase, alphanumeric with hyphens only
  4. ✓ name doesn't start/end with hyphen or have consecutive hyphens
  5. ✓ description is 1-1024 characters
  6. ✓ Body content provides clear instructions

Common Issues:

  • name doesn't match directory name
  • Uppercase characters in name
  • Description too vague (should include trigger keywords)
  • Missing instructions in body

Prompt File Validation (<provider>/prompts/*.prompt.md)

Required Structure:

---
mode: agent
description: What this prompt does
---

[Prompt template with {{variables}}]

Supported Frontmatter Attributes:

  • mode (optional) - One of: agent (default), ask, edit, generate
  • tools (optional) - Available tools for this prompt
  • description (optional but recommended) - What the prompt accomplishes

Checks:

  1. ✓ File has .prompt.md extension
  2. ✓ If mode present, it's a valid value
  3. ✓ Variables use {{variableName}} syntax
  4. ✓ Body content exists (the prompt itself)

Common Issues:

  • Wrong extension (.md instead of .prompt.md)
  • Invalid mode value
  • Undefined variables in template

Instruction File Validation (<provider>/instructions/*.instructions.md)

Required Structure:

---
applyTo: "**/*.ts"
---

[Contextual instructions]

Supported Frontmatter Attributes:

  • applyTo (required) - Glob pattern(s) for when instructions apply

Checks:

  1. ✓ File has .instructions.md extension
  2. ✓ applyTo field exists
  3. ✓ applyTo contains valid glob pattern(s)
  4. ✓ Body content provides meaningful guidance

Common Issues:

  • Wrong extension
  • Missing applyTo field
  • Invalid glob syntax
  • Empty or minimal instructions

Output Format

## Validation: [filename]

**Type:** [Agent|Skill|Prompt|Instruction]
**Status:** ✅ Valid | ⚠️ Warnings | ❌ Invalid

### Issues
- [Issue 1 with line number if applicable]
- [Issue 2]

### Recommendations
- [Suggestion for improvement]

Batch Validation

When validating all files, provide summary:

## Validation Summary

| Type | Total | Valid | Warnings | Invalid |
|------|-------|-------|----------|---------|
| Agents | X | X | X | X || Sub-Agents | X | X | X | X || Skills | X | X | X | X |
| Prompts | X | X | X | X |
| Instructions | X | X | X | X |

### Files Requiring Attention
- [List files with issues]

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.