Cloudflare's security-audit-skill turns coding agents into adversarial auditors
Cloudflare/Security-Audit-Skill
Cloudflare open-sourced a coding-agent skill that runs six-phase security audits with isolated hunters, fresh verifiers, and machine-readable findings. Each candidate is checked by an agent that didn't find it, and multiple runs are additive — in testing, one run found only about half the vulnerabilities that repeated runs uncovered. It's the single-repo seed of Cloudflare's fleet-wide vulnerability harness.
Multiple runs improve coverage. In our test runs, a single run found roughly half of the vulnerabilities that repeated runs found in total.
- gbrindisi
Shameless plug: in case someone finds this requiring too many tokens, we shared the recipe on how we built our own in house audit skill so that it can easily be replicated and tuned to different environments https://www.synthesia.io/post/automating-code-security-revie...
- prodigycorp
Hi Cloudflare people, if you are reading this. Please clean up your Cloudflare. Skills. There are way too many skills for the platform. You should consolidate all of your skills into a single skill and route everything thru that skill. The way it is right now pollutes our context window.
- wslh
Tip for security professionals using LLMs: audit skills that explicitly frame the task as security research sometimes trigger refusals from the top OpenAI and Anthropic models because they guard against misuse. What works for me: separate skills for bug classes (and bugs in general) without the security framing, plus another skill that combines their findings to spot security bugs.
- acedTrex
Incredible, a post and repo dedicated to a markdown file, the downfall of this field has been swift.
- drchaim
I threw 1M tokens for nothing in a medium codebase.