We've been working on the release of our new dashboard. We are now at the last stretch. We made multiple rounds of QA to clear out bugs before they reach production.
I saw a tweet by @jessethanley speaking about Detail[.]dev from @danlovesproofs and decided to give it a go.
It's a tool that automatically scans your repo, runs multiple agents on it, and finds bugs or security vulnerabilities in your code.
Even with four QAs working full-time on the app, it still found some very nasty bugs.
What was more surprising was that the bugs that it found needed it to really understand how proxies work, how specifically anyIP proxy works, and it was still able to understand all of it.
I'd recommend that anyone give it a try.
AI should help you ship better code.
That's why I love @detaildotdev so much. You connect it to GitHub and it surfaces bugs you missed.
And the bugs it finds are genuine.
We ran a chess tournament for bugs.
The question we wanted to answer: are bugs from Detail "important"? How do they compare to what code review bots catch?
One of the most important ways we benchmark ourselves is that we want the bugs we generate to be significantly more important than the typical comment from a code review bot.
We took a week of findings from CR bots running on OpenClaw and vLLM, plus findings from Detail on the same week of changes. We put them through an LLM-as-judge tournament.
We fed the head-to-head results into a Bradley-Terry model to compute ELO ratings for bugs. Out comes a global ranking from most to least important.
Awesome exploration from @sachiniyergreen below, with methodology, charts, and a PostHog secret exfiltration vuln that four code review bots missed.
Do this today:
Sign up to Detail[.]dev.
Connect it to Slack.
Run an audit.
It'll post vulns to the Slack channel.
Tag @cursor_ai or @OpenAI in each ticket.
Close them out.
Good Friday fun (๑•﹏•)
🚀 Just published: Detail @detaildotdev – Advanced AI scans your codebase to discover security vulnerabilities, data loss risks, and billing mistakes that traditional tools miss. High signal-to-noise ratio.
https://t.co/hTW0mQqmPt
Wild:
- ran https://t.co/ExBukH9Ytk on my repo
- it flagged a bunch of bugs and turned them into GitHub issues
- asked claude to pick them up from Glass
- fixed everything ready to be merged in minutes
Pretty cool:
- tried https://t.co/1PWJL8RvgB
- pointed it at Shiori repo
- found some bugs (and a few false positives)
- converted them all to github issues
- pointed @conductor_build at all the issues
- all bugs fixed/patched in minutes in parallel worktrees