Audits/Repository Explorer v2
Deterministic analysis · no LLM calls

Repository Explorer v2

See where AI instructions can go.

Trace prompts, skills, workflows, tool routers, MCP servers, and memory systems to the sensitive actions they can reach.

Analyze a single artifact

Paste a prompt, skill, MCP config, or agent file

Same engine, same report as a full repository scan — just one file. Pick an example or paste your own.

Analyzed as prompts/reviewer.prompt · no LLM calls

Or scan a whole repository
GitHub repositoryPublic
Scan from a GitHub URLPaste a public repo URL — the prioritized AI-relevant files are read from GitHub and scanned, no folder needed.
Public repos only for now · private repos (token) coming next.

Full local repository analysis

Runs locally with no uploads and no LLM calls.

npx @promptsonar/cli repo .