sigrid-ai-toolkit

Run Sigrid CI locally on the current working tree and return structured quality feedback. Use when the user or agent wants maintainability, security, or open source health feedback on local code without pushing to a remote or triggering a CI pipeline.

Remediate Open Source Health (OSH) findings from Sigrid — vulnerabilities, outdated dependencies, license issues, inactive/unstable libraries, and unmanaged dependencies. Produces merge requests for confident fixes, or researched GitLab issues when human judgment is needed. Use when the user wants to fix OSH findings, resolve dependency vulnerabilities, update outdated libraries, address license risks, or act on any Sigrid open source health result. Also trigger when the user mentions "fix dependencies", "update vulnerable packages", "OSH remediation", "CVE", "SBOM", "dependency risk", or wants to act on open source risks surfaced by Sigrid.

Guided maintainability diagnosis for a Sigrid system. Finds the weakest property, explains what is causing it, and surfaces the top refactoring candidates. Use when the user wants to understand where to start improving code quality, or asks what their biggest maintainability problem is.

Use when the user wants to act on a Sigrid maintainability diagnosis — fix refactoring candidates, reduce a low-rated property, or apply improvements surfaced by sigrid-diagnose. Trigger on phrases like "fix this", "apply these improvements", "start refactoring", "work on the candidates", or any request to make code changes based on Sigrid findings. Makes actual code changes. Run after sigrid-diagnose, or run sigrid-diagnose first if no diagnosis is in context yet.