Matt Pocock says his /code-review leans on project specs, not generic heuristics
mattpocockuk · x · 2026-07-25
The author clarifies that their /code-review setup does use some generic review logic, mainly Fowler-style refactoring heuristics.
But the important part is that most of the review is delegated to the project’s own coding standards and spec compliance checks, reinforcing the idea that useful review agents need project-specific context rather than a one-size-fits-all prompt.
Related event: Effective AI Code Review Relies on Project-Specific Standards(2 posts)→
More from coding & agent
- He benchmarks new models by having them build their own copilot — $28 of compute — bucolucas · 2026-10-27
- Should agents be allowed to change plans mid-task? Reddit debates the drift boundary — wistfulcracker84 · 2026-10-27
- HeyGen + Canva MCP turns slide speaker notes into an avatar presentation video — HeyGen · 2026-10-27
- Local beats cloud: Qwen 3.8 FN handles private student data as well as frontier models — dr_lm · 2026-10-27
- LLM Tails cuts agent input tokens 80% by making agents read before they write — savadks1818 · 2026-10-27
- 10 practical ways to run Astra computer use on Premiere Pro for video editing — beechinour · 2026-10-27