search

I took a look at your search skill and wanted to share some thoughts.

Links:

• GitHub
• Your agentic skill in the SkillzWave marketplace

The TL;DR

You're at 76/100, solidly in C territory. This is based on Anthropic's best practices for Claude skills. Your strongest area is Spec Compliance (13/15) — the frontmatter and naming are tight. Weakest area is Progressive Disclosure Architecture (18/30) — you've got a 175-line file with no table of contents and some content that could be layered better.

What's Working Well

• Spec compliance is locked in — Valid YAML, correct hyphen-case naming, solid third-person description with specific trigger phrases covering local and web search scenarios.
• Clear 3-step workflow — The mission-based input model is straightforward, and your output schema is complete and usable.
• Strong trigger phrases — You nailed discoverability with phrases like "search locally", "web search", "find information" — Claude will know exactly when to call this.
• Concrete examples — Three mission examples help people understand how to use it, which bumped up the utility score.

The Big One: Missing Error Handling Guidance

This is your biggest miss. Your "Important Rules" section covers the happy path only — you've got nothing about handling empty results, rate limits, or failed searches. That's a high-severity gap because Claude needs to know how to handle failures gracefully.

Fix: Add an Error Handling section with specific schemas and retry logic. Something like:

## Error Handling

**Empty Results:** Return schema with `results: []` and `error: "No results found for [mission]"`
**Rate Limiting:** Retry logic for external providers (jina.ai, firecrawl)
**Malformed Mission:** Va...