[Due for payment 2025-12-07] [$250] [AI Reviewer] PERF-6 rule is not correctly understood by agent

[kacper-mikolajczak]

Problem

PERF-6 rules is triggered many times in scenarios it should not be, like reported here by Vit #72663 (comment).

Solution

Research how rule could be re-phrases so it does not alert on false-positives.

Lint rule PR: Expensify/eslint-config-expensify#170

App PR: #76019

Upwork Automation - Do Not Edit

• Upwork Job URL: https://www.upwork.com/jobs/~021993665608221777255
• Upwork Job ID: 1993665608221777255
• Last Price Increase: 2025-11-26

Issue Owner

Current Issue Owner: @

Issue Owner

Current Issue Owner: @greg-schroeder

[kacper-mikolajczak]

After checking in more detail how PERF-6 rule operates, I came to conclusion it should not be really a responsibility of ai reviewer to detect.

I am convinced we could, with really high confidence, guard against it with regular static analysis via linter rule.

I am going to draft a PoC for that soon.

[kacper-mikolajczak]

Created linter PR: Expensify/eslint-config-expensify#170

Next steps:

1. Discuss PoC with the team
2. Test new linter rule on existing PRs
3. Remove PERF-6 rule if linter works correctly

[kacper-mikolajczak]

Update: Discussion is progress