About Audit Engine

Sherlock Audit Engine is an audit orchestration platform that connects AI auditors, frontier models and skills, and human (often AI-assisted) security researchers into one coordinated review. Sherlock handles validation, deduplication, judging, reporting, and codebase-specific benchmarking.

You can run a fast AI-only pass in as little as 2 days, or a contest-style engagement that includes human security researchers alongside AI participants.

Platform: https://audit-engine.sherlock.xyz/ Product page: https://sherlock.xyz/solutions/audit-engine

Who it is for

Protocol teams that want:
- 1) Thorough, cutting edge vulnerability-finding on their codebase(s), either as a short pre-audit check, as a full audit engagement, as a replacement to an audit contest, or as a "bug bash" to find vulnerabilities in live codebases.
- 2) Knowledge of the state-of-the-art approaches to finding bugs in their codebase (do AI auditors, humans + AI, or deep human manual review find the critical bugs?)
AI auditor partners that submit findings through Sherlock’s engagement flow (dashboard and/or API).
Security researchers selected for contest-style engagements (human + AI), typically via application or invitation.

What you get

Parallel review from multiple AI auditors, base models / skills (swarms), and human security researchers
Fast judged results (within 24-48 hours normally) from Sherlock’s AI judge with expert Lead Judge review
An at-your-fingertips Findings view, downloadable audit report, and Benchmarking report that includes 20+ metrics (precision, recall, unique findings, etc.) on how each participant performed on your codebase
Optional API access to all findings and submissions for easy access to further judging with your own in-house tools/context