bowtiedswan/solodit-api-skill
Overview
This skill provides fast search and retrieval across 50,000+ smart contract security findings aggregated by Cyfrin Solodit. It exposes eight MCP tools with intelligent caching so you can search, filter, and fetch detailed vulnerability reports without repeatedly hitting the upstream API. The skill is optimized for audit preparation, vulnerability research, and threat modeling for blockchain protocols.
How this skill works
The skill queries the Solodit database using parameters like keywords, severity, audit firm, tags, protocol category, language, and quality score. Results are cached at multiple levels (search: 5 minutes, findings: 1 hour, metadata: 24 hours) to reduce rate-limit impact and speed repeated queries. Additional tools list metadata (audit firms, tags, languages), provide database statistics, and let you clear cache for fresh results.
When to use it
- Preparing security audits and creating targeted test cases for smart contracts
- Researching historical vulnerabilities by pattern, tag, or audit firm
- Identifying high-quality or recent findings for threat modeling
- Exploring common attack types across protocol categories (DeFi, NFT, Bridges)
- Filtering findings by language (Solidity, Rust, Cairo) or protocol name
Best practices
- Start with narrow filters (impact, tags, protocol) to reduce noise and cost
- Use sort options (recency, quality, rarity) to prioritize high-value findings
- Leverage cached metadata (audit firms, tags) to build consistent filters and UIs
- Respect rate limits: prefer cached results and batch queries when possible
- Use clear_cache only when you need fresh metadata or recent finding updates
Example use cases
- Search all HIGH severity lending protocol findings from the last 90 days sorted by quality score
- Find oracle manipulation examples in DeFi with quality >= 4 for vulnerability patterns
- List available audit firms and tags to populate a security dashboard
- Retrieve a specific finding by ID or slug to include as a mitigation reference in a report
- Investigate flash loan attack patterns by searching tag "Flash Loan" sorted by rarity
FAQ
The database contains over 50,000 aggregated findings from many top audit firms.
How does caching affect search freshness?
Search results are cached for 5 minutes, individual findings for 1 hour, and metadata for 24 hours. Use the clear_cache tool to force fresh data when needed.