Deep Audits
A deep audit runs the complete engagement pipeline: baseline scanning, asset discovery, and then the Golem agent’s full attack chain against every live asset.What it covers
Deep mode runs the full methodology:- Baseline scan — nuclei, security headers, common paths (deterministic)
- Asset discovery — passive subdomain enumeration from nine OSINT sources, then liveness validation
- Live probing — HTTP probing and tech-stack fingerprinting of every live asset
- Agent assessment — Golem plans and executes its own attack chain: crawling, content discovery, injection testing, auth attacks, stack-specific exploitation
- Finding validation — every agent claim is checked against real command output before it’s accepted
- Reporting — score with breakdown, coverage summary, Markdown report
Security focus
Deep audits consume your app’s security focus areas — the selected emphases are injected into the agent’s system prompt. You can also override the focus per run. With no focus set, the agent covers the standard breadth.What makes it different from Shallow
Running a deep audit
Dashboard- Open the app
- Click Launch Audit
- Select Deep
- Optionally adjust the security focus for this run
- Launch
Expected output
A typical deep audit produces 5–40 findings depending on surface area, including High and Critical findings for confirmed RCE, auth bypass, or sensitive data exposure — each with evidence traced to a real command execution. The audit also records a coverage summary: every phase and technique attempted, so you can see not just what was found but what was actually tested.Next steps
Autonomous Audits
Add your own mission brief.
Methodology
Read the phased chain in detail.