vigolium-audit

To automate a security audit, run a multi-agent pipeline executing intel, SAST, adversarial review, and PoC generation to produce structured findings with reduced false positives. This workflow autonomously triages your codebase and outputs a finalized report without manual coordination.

Yes, you can get JSON output from a codebase vulnerability assessment by using the optional --json flag to emit NDJSON streams. This machine-readable format allows direct integration of finalized security findings and PoC data into CI/CD pipelines.

The best way to resume an interrupted codebase vulnerability audit is to use the revisit or reinvest modes. These modes continue interrupted runs, allowing the multi-agent pipeline to pick up where it left off and finalize structured findings without losing prior analysis.

Multi-agent SAST triage works by passing static analysis results through adversarial review and PoC generation to validate exploitable vulnerabilities. This sequential pipeline filters out false positives and produces a stable findings tree with confirmed security issues.

Yes, vigolium-audit requires compatible Claude Code or Codex tooling installed on your PATH to run a deep security audit. You must install the vigolium-audit binary and invoke it with the correct mode, agent, and target parameters to execute the pipeline.

Use lite mode for faster repository security analysis and deep mode for thorough assessments. Deep mode runs the full intel-to-reporting pipeline for critical vulnerabilities, while lite mode matches time constraints when you need a quicker codebase triage.