marian2js

marian2js / skill-doctor

Public

Help your agents create better skills

agent-skills agents agentskills claude-skills doctor
18
1
100% credibility
Found Mar 15, 2026 at 18 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Skill Doctor scans folders of AI agent skills for quality issues in documentation, resources, triggers, instructions, and evaluations, delivering a 0-100 score with actionable diagnostics.

How It Works

1
📖 Discover Skill Doctor

You learn about a friendly tool that checks your AI agent skills for problems before they cause issues.

2
📁 Point to your folder

You simply tell the tool to look at the folder holding your skills with one easy command.

3
🔍 Automatic checkup

It quickly scans all your skills for missing details, broken links, unclear instructions, and more.

4
📊 Get your score

You see a clear 0-100 score for each skill plus a summary of what's great and what to fix.

5
💡 Follow friendly tips

Simple advice shows exactly where issues are and how to make your skills stronger.

✅ Healthy skills ready

Your skills score perfectly and work smoothly with AI agents without surprises.

Sign up to see the full architecture

4 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 18 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is skill-doctor?

Skill-doctor scans agent skill directories for static issues in SKILL.md files, checking frontmatter validity, broken local links, weak trigger descriptions, missing workflow guidance, and malformed evals/evals.json. It delivers a conservative 0-100 score with actionable diagnostics, helping AI agents avoid routing failures, onboarding snags, or eval breaks. Built in TypeScript, it runs via npx skill-doctor ., outputs JSON for scripting, or plugs into GitHub Actions for CI.

Why is it gaining traction?

Unlike generic linters, it targets agent skills specifically—calibrated on real-world corpora like Anthropic's—with strictness levels (default, strict, pedantic) to match your tolerance for style nits. Developers grab it for the instant workspace overview table and per-skill breakdowns, spotting bundle integrity gaps that generic YAML/MD checkers miss. AI agents help a lot in developing an understanding of tasks, and this ensures skills trigger reliably without vague generics.

Who should use this?

Agent skill authors tuning descriptions for better routing, teams building evals for benchmarks, or ops folks enforcing quality in monorepos via GitHub Actions. Ideal for Anthropic/Firety users where help agents prioritize cases effectively by tracking milestone time, or anyone debugging why skills flake in production.

Verdict

Grab it now if you're shipping skills—npx simplicity and focused rules deliver real value fast. With 18 stars and 1.0% credibility score, it's early but docs are crisp and tests solid; fork if needed, but watch for ecosystem adoption.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.