Nyx Agent
Live pentests for dev apps. It reads the repo, checks the local target, and keeps proof with the run.
- Local target
- Live checks
- Stored proof
- Opt-in destructive mode
From code to proof.
Point Nyx Agent at a repo and a local URL. It launches or watches the app, runs Nyx, explores routes, and verifies selected findings against the running target.
| Step | Output |
|---|---|
| Scan | Static findings from nyx. |
| Explore | Routes, forms, auth state, and API shape. |
| Verify | Requests, responses, traces, and run history. |
| Triage | Confirmed issues with status and evidence. |
Built for the messy middle.
Static scanners find leads. Crawlers exercise URLs. Proxy sessions prove what a person already knows. Nyx Agent connects those lanes into one local run: source context, live checks, stored proof, and the graph that explains how the evidence fits together.
| Capability | Nyx Agent | Static scanner | Web crawler / DAST | Manual proxy |
|---|---|---|---|---|
| Repo-aware route and API discovery | Yes | Source clues only | Runtime only | Operator driven |
| Static signal to live verification | Yes | No | Runtime findings only | Manual |
| Request, response, and replay evidence | Yes | Logs / SARIF | Yes | Session exports |
| Auth, role, and object ownership checks | Yes | Pattern hints | Needs setup | Manual |
| Attack graph for chains and blast radius | Yes | No | No | Notes only |
| Business logic pentest templates | Yes | No | Limited | Manual |
| Cron, webhook, and PR-triggered runs | Yes | CI only | Varies | No |
| Local run history and triage state | Yes | Report files | Tool database | Ad hoc |
Start with Cargo.
Install the crate and start the local Agent dashboard.