Crawler policy

OpenSkillIndex operates a public-interest crawler that discovers agentic AI skills from public sources.

User agent

OpenSkillIndexBot/1.0 (+https://openskillindex.com/bot)

What we fetch

• Files named SKILL.md, SKILL.yaml, SKILL.yml from public GitHub repositories
• The repository's public metadata (stars, license, homepage, last push) to annotate the index
• Publicly linked skill manifests from opt-in marketplaces (currently: none — all scrape-first crawling is disabled)

Rate limits we respect

• GitHub's published rate limits (authenticated, single app token)
• Exponential backoff on any 4xx/5xx response
• Per-source kill switch: any source can be instantly disabled

We do not

• Scrape private repositories or anything behind authentication
• Scrape marketplaces whose terms of service do not permit automated access
• Re-host code — every install link points back to the original source
• Sell placement or ranking — the formula is open source

Opt out / takedowns

Email support@openskillindex.com with the repository or skill URL. We hide public records within 24 hours. Retention of audit data follows our neutrality commitments.