Case study · Offensive security agent¶
You are here: Case study · Offensive security agent · Power user
Next: Playbooks
Antigen, as announced: an agent that maps a surface, probes it, ships fixes for review, and tests again.
| Path | Power user |
| Source | @antigenco |
| Posted | 2026-10-08, @antigenco |
| Concept | Find, review, fix, retest |
The situation¶
The post is a product introduction. The verbs are map, exploit, secure, probe, ship fixes for the team's review, and test again. The post does not include a resolved documentation URL.
Claims¶
| Claim | Status |
|---|---|
| Antigen is an offensive security agent | Per post |
| It maps and probes an attack surface, then proposes fixes | Per post |
| Fixes are shipped for a team's review | Per post. This is the part the concept page treats as the gate |
| A retest checks that the issues were patched | Per post |
| How probing or exploitation is done | Not taught. Out of scope for this guide |
What to steal¶
The loop with a person on the merge, and a retest that does not trust the agent's summary of itself. Record findings as pages so the next run sees status, not a fresh scan with no memory.
What to leave¶
Any request to expand "exploit" into steps. Any use against a system you are not allowed to test. The security index states that boundary for later sources too.
Sources¶
Continue¶
- Playbooks
- Previous: Local inference
Map | Paths | Glossary | References