Skip to content

fix: refine review-pr skill with evidence-backed rules from reviewer feedback#25

Draft
captainsafia wants to merge 1 commit into
mainfrom
fix/review-pr-skill-feedback-2026-06-05
Draft

fix: refine review-pr skill with evidence-backed rules from reviewer feedback#25
captainsafia wants to merge 1 commit into
mainfrom
fix/review-pr-skill-feedback-2026-06-05

Conversation

@captainsafia
Copy link
Copy Markdown
Contributor

Summary

Adds three new rules to the review-pr skill's Review Scope section, each backed by evidence from human reviewer feedback on warpdotdev/warp-server PRs in the week of 2026-05-29 to 2026-06-05.

New rules

  1. Verify symbols in full file context — Before claiming a symbol/function/pattern is missing, check the full file, not just the diff. (Evidence: PR #11601 — reviewer corrected bot for claiming a symbol didn't exist when it was in the same file.)

  2. Calibrate severity for optimization code — When reviewing optimization or heuristic code (caches, early exits, approximate checks), calibrate severity to the tolerance of the use case. Don't treat imperfect optimizations as correctness bugs. (Evidence: PR #11624 — reviewer noted the bot missed the nuance that the check was "just an optimization" and didn't need to be perfect.)

  3. Don't suggest infrastructure-dependent resources — Don't suggest switching to fields/APIs/resources that depend on external infrastructure (e.g. Terraform) without confirming they're available. (Evidence: PR #11668 — reviewer corrected bot for suggesting a field not yet provisioned via Terraform.)

What was not changed

  • review-spec skill: Only one spec PR (#11665) had human comments this week, and they were clarifying Q&A between the spec author and reviewers — no actionable signal for skill changes.
  • No feedback_items (direct replies to bot review comments) were found this week, but several human review comment threads included substantive corrections to bot behavior.

cc @captainsafia

Conversation: https://staging.warp.dev/conversation/6f40db5f-b328-4025-8449-6460076103c3
Run: https://oz.staging.warp.dev/runs/019e984c-681e-72bd-9693-f349a3ad1438

This PR was generated with Oz.

Co-Authored-By: Oz oz-agent@warp.dev

…feedback

Add three new rules to the Review Scope section based on patterns
observed in human reviewer feedback from the week of 2026-05-29
to 2026-06-05 across warpdotdev/warp-server PRs:

1. Verify symbols exist in full file context, not just the diff
   (PR #11601 — reviewer corrected bot for claiming a symbol was
   missing when it existed in the same file).

2. Calibrate severity for optimization/heuristic code vs correctness
   code (PR #11624 — reviewer noted the bot treated an optimization
   imperfection as a correctness bug).

3. Do not suggest infrastructure-dependent resources without confirming
   availability (PR #11668 — reviewer corrected bot for suggesting a
   field not yet provisioned via Terraform).

No changes to review-spec: only one spec PR had human comments this
week and they were clarifying Q&A with no actionable signal.

Co-Authored-By: Oz <oz-agent@warp.dev>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant