SCAN. TEST. SURVIVE.
The
Gauntlet.
REAL SECURITY. REAL TESTING. REAL BUILDERS.
A release-blocking security orchestrator. It scans contracts, frontend, backend and deployment code, scores the result, and blocks anything unsafe or incomplete.

Good code
survives.
survives.
Scan status
● Published scan indexLoading…Verified verdicts appear here from the scan index. Public reports are summary-only.
Gauntlet flow.
01
SUBMIT
Project enters the Gauntlet with gauntlet.yml, threat model and invariants.
02
SCAN
Slither, Echidna, Semgrep, Gitleaks, Foundry tests, npm/pip audit and built-in checks.
03
VERDICT
Findings normalized, scored and graded. Missing required tools fail closed.
04
HERMES
Blocked builds get a fix prompt. Patch the root cause, rerun the full scan.
05
SURVIVE
Passing builds move on to human and independent review.
↺ If security or tests fail, it returns to Hermes for fixes — max 3 rounds, then blocked. A scanner that can't run is an infrastructure issue, not a code finding.
Verdicts
✕
BLOCKEDExploit, failed required check, missing required tool, or score below 85.
◐
STATIC_ONLYStatic review of untrusted code. Never a launch approval.
◇
TESTNET_ONLYChecks passed, but the risk tier requires independent review.
!
REVIEW_REQUIREDNo automatic blocker, but score is below the pilot threshold.
✓
CAPPED_PILOT_CANDIDATENo blocker found. Launch limited to the stated cap and exact reviewed release.
A
95–100
B
85–94
C
70–84
D
50–69
F
0–49
A high score never overrides a Critical or High finding. A pass applies only to the exact commit and configuration scanned — it is not a safety certificate or formal audit.
