CtrlK
BlogDocsLog inGet started
Tessl Logo

sentry-backend-bugs

Review Sentry Python and Django changes for bug patterns drawn from real production issues. Use when reviewing a backend diff or PR, checking Warden findings, auditing the current branch, reviewing production-error patterns, or looking for common regressions in `src/` and `tests/`.

87

0.97x
Quality

83%

Does it follow best practices?

Impact

94%

0.97x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No known issues

SKILL.md
Quality
Evals
Security

Evaluation results

82%

-18%

Django ORM Safety Review

ORM .get() bug vs. intentional-crash discrimination

Criteria
Without context
With context

Genuine bug flagged

100%

100%

Async task bug flagged

100%

100%

Infrastructure invariant NOT flagged

100%

100%

Parent-validated endpoint NOT flagged

100%

0%

Correct HTTP codes in fixes

100%

100%

No comment-only fixes

100%

100%

Confidence levels assigned

100%

100%

Triggering input specified

100%

100%

filter().first() fix pattern

100%

0%

No invented findings

100%

100%

100%

Webhook Handler Code Review

Webhook handler: header/JSON/unpack bug review

Criteria
Without context
With context

Header KeyError flagged

100%

100%

Header fix uses .get()

100%

100%

JSON parse error flagged

100%

100%

JSON fix wraps in try/except

100%

100%

Tuple unpack flagged

100%

100%

Tuple unpack fix validates length

100%

100%

Handler dispatch bug flagged

100%

100%

Confidence levels stated

100%

100%

Triggering inputs specified

100%

100%

100%

10%

Background Worker Code Review

Integer overflow and concurrent dict mutation

Criteria
Without context
With context

Integer overflow flagged

100%

100%

Overflow fix caps at 2_147_483_647

100%

100%

Overflow fix uses Least() or equivalent

0%

100%

Dict mutation bug flagged

100%

100%

Dict fix avoids mutating during iteration

100%

100%

Concurrency concern noted for publisher

100%

100%

Confidence levels stated

100%

100%

Triggering state described

100%

100%

Fixes include actual code

100%

100%

Repository
getsentry/sentry
Evaluated
Agent
Claude Code
Model
Claude Sonnet 4.6

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.