ClawSearch CLI

Safe skill discovery for AI agents. Search 33,000+ skills with Trust Score, 10-language support, and @author lookup. Continuous auditing ~2,000/day.

Install

npm install -g clawsearch

Or use directly with npx:

npx clawsearch slack

Usage

# Search skills (10 languages supported)
clawsearch slack
clawsearch "消息推送" --safe
clawsearch tool --type tool --limit 10

# Search by author
clawsearch @steipete
clawsearch @steipete --json

# View Trust Score breakdown
clawsearch trust slack

# Compare skills side by side
clawsearch compare slack discord feishu

# Pre-install security check
clawsearch check slack
clawsearch check https://clawhub.ai/steipete/slack

# Show trending skills
clawsearch trending

# JSON output (for agent integration)
clawsearch security --json
clawsearch trust slack --json

Options

Option	Description
--safe	Only show SAFE skills
--json	Output as JSON (for agent integration)
--type <type>	Filter by type (tool, prompt_template, mcp_server, etc.)
--limit <n>	Max results (default 20)

Trust Score

Every skill gets a Trust Score (0-1) based on four dimensions:

• Safety (40%) — ClawSec 5-tier audit verdict + D2 security score
• Author (20%) — GitHub account age, skill count, malicious ratio
• Community (20%) — Downloads, stars
• Transparency (20%) — Open source, license, no obfuscation

●●●●● Verified Safe    (0.8-1.0)
●●●●○ Mostly Safe      (0.6-0.8)
●●●○○ Use with Caution (0.4-0.6)
●●○○○ Suspicious       (0.2-0.4)
●○○○○ Dangerous        (0.0-0.2)

Supported Languages

Search in: English, 中文, 日本語, 한국어, Deutsch, Français, Español, Português, Русский, العربية

API

Powered by ClawSearch — the security-first AI agent skill search engine.

• Web: https://clawsearch.cc
• API: https://api.clawsearch.cc
• API Docs: https://api.clawsearch.cc/api/v1/docs
• Audit Engine: ClawSec
• Pre-install Guard: clawsearch-guard

License

MIT