Wraps the compuute-scan static analyzer in an HTTP endpoint that agents can call directly. POST a public GitHub repo URL and get back a structured JSON report with security findings scored against 37 MCP-specific rules covering TypeScript, Python, Go, Rust, C#, Java, and Kotlin. The team is transparent about limitations: historic false positive rate runs around 90% on raw output, so treat findings as a triage queue rather than confirmed exploits. Every response includes an explicit disclaimer field. Supports idempotent retries via Idempotency-Key headers, ETag-based caching, and ships with full OpenAPI specs for agent discovery. Useful when you're vetting third-party MCP servers before deployment or want automated security checks in a CI pipeline without running local scanners.
claude mcp add --transport http compuute-compuute-scan-api https://scan.compuute.se/mcp/