Context
During the PR review cycle for #10242, the reviewer (Claude) flagged a retry loop as a violation of the "surface, don't obscure" pattern. The author (Gemini) responded by running an empirical binary isolation test—disabling the retry and rebooting the harness—which proved in 5 minutes that the busy_timeout pragma alone fixed the underlying race.
This converted an architectural argument into a clean empirical result rapidly and respectfully.
The Problem
The current PR and PR Review skill guidelines don't explicitly highlight this collaborative "prove it with an isolation test" pattern as an ideal resolution path for architectural disputes.
The Architectural Reality
Agents operate in a fast-feedback environment. When a reviewer flags a pattern violation, the author can either (a) document why the pattern is necessary, or (b) run an isolation test to show the pattern is dead weight. Option (b) leverages the unique strengths of the local agent harness.
The Fix
Extend the pull-request and pr-review skills to formalize the "Isolation-Test-After-Review" pattern.
Acceptance Criteria
Out of Scope
- Modifying the GitHub Actions workflows or PR templates.
Origin Session ID
Origin Session ID: 1c5933cc-a2d0-4296-ae3f-f4e815d385a2
Context
During the PR review cycle for #10242, the reviewer (Claude) flagged a retry loop as a violation of the "surface, don't obscure" pattern. The author (Gemini) responded by running an empirical binary isolation test—disabling the retry and rebooting the harness—which proved in 5 minutes that the
busy_timeoutpragma alone fixed the underlying race.This converted an architectural argument into a clean empirical result rapidly and respectfully.
The Problem
The current PR and PR Review skill guidelines don't explicitly highlight this collaborative "prove it with an isolation test" pattern as an ideal resolution path for architectural disputes.
The Architectural Reality
Agents operate in a fast-feedback environment. When a reviewer flags a pattern violation, the author can either (a) document why the pattern is necessary, or (b) run an isolation test to show the pattern is dead weight. Option (b) leverages the unique strengths of the local agent harness.
The Fix
Extend the
pull-requestandpr-reviewskills to formalize the "Isolation-Test-After-Review" pattern.Acceptance Criteria
.agent/skills/pull-request/SKILL.mdas an ideal author response strategy..agent/skills/pr-review/SKILL.mdto guide reviewers in suggesting isolation tests when challenging patterns.Out of Scope
Origin Session ID
Origin Session ID: 1c5933cc-a2d0-4296-ae3f-f4e815d385a2