Skip to content

Agent Skill Validator

Validate skill definitions across OpenClaw, Claude, Codex, and MCP with portability scoring and exact fixes

Agent Skill Validator

Portability score 100, 0 errors, 0 warnings

Portability score
100
Errors
0
Warnings
0
No portability issues detected for selected runtimes.

Validate first, convert next:SkillSpec Converter

What this tool does

Agent Skill Validator checks a skill or prompt definition for OpenClaw, Claude, Codex, and MCP compatibility, flagging missing fields, ambiguous triggers, unsafe instructions, and portability breaks. It runs in your browser, scores how cleanly the skill ports to each platform, and returns the exact edits needed to fix every issue it finds.

Updated . Provided as is. Check the output before you rely on it in production.

How to use Agent Skill Validator

  1. 1

    Paste your skill definition

    Add your skill spec in JSON, YAML, or markdown. Include name, description, trigger, instructions, and tools for the best validation coverage.

  2. 2

    Select target runtimes

    Choose OpenClaw, Claude, Codex, and/or MCP. The validator applies runtime-specific lint rules and portability checks per target.

  3. 3

    Review errors and warnings

    Inspect diagnostics sorted by severity. High-severity errors indicate portability blockers, while warnings highlight clarity and compatibility risks.

  4. 4

    Apply suggested fixes

    Use the Suggested Fixes tab to address missing fields, ambiguous triggers, naming mismatches, and schema gaps quickly.

  5. 5

    Copy the patch preview and revalidate

    Copy the generated patch preview, apply the changes to your canonical skill, then run validation again until portability score is acceptable.

Questions and answers

What does Agent Skill Validator check?
A skill definition tells an agent what a skill does, when to use it, what steps to follow and which tools it may call. This validator parses JSON, YAML or Markdown skills, checks name, description, trigger, steps, unsafe instructions and MCP inputSchema, and scores portability 0-100 for OpenClaw, Claude, Codex and MCP.
Which input formats are supported?
You can paste skill definitions as JSON, YAML, or markdown-style skill docs. The validator normalizes structure and runs deterministic lint rules client-side in your browser.
Does Agent Skill Validator execute my tools or prompts?
No. It performs static text analysis only. No tool execution, no remote calls, and no server-side processing of your skill content.
What is the portability score?
The portability score (0-100) summarizes how safely your skill can be reused across selected runtimes. High-severity issues reduce score more than medium or low warnings.
Can I apply fixes automatically?
The tool provides a patch preview with suggested normalized fields and safer defaults. Copy that patch, apply it to your canonical skill file, and revalidate.
For AI agents: how to call this tool

Machine-readable contract, endpoints and examples. Humans can ignore this section.

Best Path For Builders

Browser workflow

Runs instantly in the browser with private local processing and copy/export-ready output.

Browser Workflow

This tool is optimized for instant in-browser execution with local data handling. Run it here and copy/export the output directly.

/agent-skill-validator/

For automation planning, fetch the canonical contract at /api/tool/agent-skill-validator.json.