Public Wake Arena AI reports

This repository collects Wake Arena AI reports that have been published by our clients.

• Request access: Wake Arena
• Looking for manual audit reports instead? See Ackee-Blockchain/public-audit-reports.

About Wake Arena

Wake Arena is a multi-agent AI security system with graph-driven reasoning and LLM-tailored static analysis detectors to help teams catch vulnerabilities before a premium audit.

Built by the auditors behind Lido, Aave, Axelar, Safe and more (Ackee Blockchain Security). Wake Arena is powered by the Wake Framework.

Why teams use it

• Pre-audit preparation: catch surface-level issues early and focus human auditors on deeper protocol logic.
• Continuous security: scan every PR, upgrade, or release to reduce security debt.
• Third-party integration: assess risk in external Solidity code before integrating tokens, oracles, or dependencies.

How it works

1. Upload/connect a Solidity codebase (or target a branch/repo).
2. Run the AI-driven audit pipeline: detector signals + multi-step reasoning guided by dependency graphs.
3. Get a report with findings, severity, and remediation guidance.
4. Fix, then audit (arrive to a manual audit with cleaner code).

Proven results

Wake Arena combines 108 detectors with deep reasoning to cover issues that pattern-based tooling often misses. On the Wake Arena landing page you can find benchmark breakdowns and reported metrics, including:

• Detected 43 / 94 high-severity issues across historical audit competition benchmarks
• 33% of all reported findings and 50% of critical findings discovered (across benchmarks and production audits)

For details, see the launch blog post.

Report contents

Each audit report typically includes:

• The findings classification used throughout the report
• Executive summary highlighting the key insights
• A list of findings (with severity and remediation guidance)

Reports are organized into folders within this repository (layout may vary over time).

Notes & disclaimers

• Reports reflect the state of the codebase at the time of the audit and within the agreed scope.
• Security is not a one-time event—treat reports as one input into an ongoing security process.

Repository contents

• Reports are stored by year (e.g. 2025/, 2026/)
• Each report PDF has a matching signature file: .pdf.sig

Verifying report signatures

Reports have a signature that can be used to verify that the PDF has not been modified when downloaded from a source other than this repository.

• Download the public.key file from this repository.
• Import it with gpg and verify:

gpg --import public.key
gpg --verify <report>.pdf.sig <report>.pdf