vuln-analysis-agent

Attack chain mapping correlates scattered vulnerability findings into multi-step exploitation paths. It relates individual issues like LFI or IDOR to show how they combine into privilege escalation or remote code execution scenarios for stakeholder reporting.

Read findings from workspace/findings/*.json and correlate them into a structured attack_chains.json file. This assigns chain severity and ordered exploitation steps, transforming raw data into a validated narrative.

Run python scripts/verify_findings.py <workspace_dir> to re-verify confirmed issues. This script emits a verification.json file, ensuring report-ready confidence for the synthesized attack chains.

Yes, cross-agent attack chain analysis builds structured JSON by correlating findings across multiple penetration agents. It assigns chain severity and ordered exploitation steps for coherent security assessment outcomes.

Yes, the process outputs a valid syntax JSON attack_chains.json file matching a required schema. It strictly enforces this structure during Phase 5.5 synthesis to ensure reliable downstream parsing.

Evidence re-verification is restricted to running verify_findings.py which emits only verification.json. It applies specifically to Phase 5.6 for web/API penetration testing workspaces and does not perform new exploitation.