discovery

You can automatically detect hardcoded secrets and dependency vulnerabilities by running automated discovery probes. The skill detects project stack markers, executes parallel probes, verifies findings by re-reading reported file locations, and assigns confidence scores to filter results by severity.

The best way to triage code quality regressions is through interactive triage of verified findings. After automated probes execute and deduplicate results, the skill formats and optionally creates VCS issues with appropriate labels and templates for team remediation.

Yes, automated codebase discovery works with JavaScript, TypeScript, Python, Docker, CI, and UI frameworks. It automatically detects project stack markers and activates relevant probes to surface findings like test gaps, accessibility issues, and design drift across those environments.

Evidence-first verification for security findings works by running grep and CLI probes, then re-reading reported file locations to confirm the issues. This process assigns confidence scores, deduplicates results, and filters by configured thresholds to reduce false positives during triage.

Yes, you can filter automated discovery findings by severity and confidence thresholds. The skill respects configured thresholds and bootstrap gates during probe execution, ensuring only verified findings with sufficient confidence scores surface for interactive triage and optional VCS issue creation.

Automated probes surface issues across code, infra, UI, architecture, and session scopes. Findings include hardcoded secrets, dependency vulnerabilities, test gaps, accessibility issues, and design drift, all verified through stack-aware detection and evidence-first confirmation.