Senior Board Audit: a four-agent adversarial code review skill (MIT, looking for feedback) #1174
mahdyhasan
started this conversation in
Show and tell
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
-
Hi all, sharing a skill I built and use on my own projects in case it's useful to others here. Feedback welcome.
What it does
senior-board-auditruns a four-agent adversarial audit on a codebase instead of a single-agent review:The audit adapts to project stage. Pre-coding asks architecture and scale questions. Mid-project hunts bottlenecks, race conditions, and resilience gaps. Pre-ship stress-tests rollback paths, load capacity, and secret rotation.
What it produces
Five files in
docs/audit/{date}/:report.md(executive summary + findings by severity)inventory.mddependency-map.mdchallenge-rounds.md(side-by-side alternatives for top 5 issues)unanswered-questions.mdIt refuses to modify any source file. Verdicts only, no patches.
Why I built it
Most reviews end too polite (rubber stamp) or too unstructured (50 nits with no ranking). The board structure forces disagreement onto the record and requires a verdict on every component. Borrows from how surgeons, pilots, and financial auditors handle high-stakes review.
Repo
https://github.com/ScaleWithAugmex/senior-board-audit
MIT licensed. Works with Claude Opus and other coding agents that can read files and reason across multiple passes. The README has install instructions for Claude Code, claude.ai, and any agent that accepts markdown-based skills.
What I'd love feedback on
Happy to answer questions or take PRs.
Beta Was this translation helpful? Give feedback.
All reactions