Automated code review for pull requests using multiple specialized agents
69
55%
Does it follow best practices?
Impact
99%
1.50xAverage score across 3 eval scenarios
Advisory
Suggest reviewing before use
Optimize this skill with Tessl
npx tessl skill review --optimize ./.claude/skills/code-review/SKILL.mdHigh-signal issue filtering
Flags delivery window bug
100%
100%
Flags unimplemented function
100%
100%
No style flag
100%
100%
No speculative flag
100%
100%
Suggestion block for simple fix
100%
100%
No suggestion block for throw
100%
100%
No duplicate comments
100%
100%
No uncertain language
100%
100%
Exactly two issues
100%
100%
File attribution
100%
100%
Comment format and CLAUDE.md scoping
Full SHA in links
100%
100%
Correct link format
100%
100%
Correct repo name in links
100%
100%
Context lines in range
0%
100%
CLAUDE.md link in violation comment
100%
100%
Suggestion block for small fix
0%
100%
No suggestion block for complex fix
100%
100%
One comment per issue
100%
100%
Path-scoped CLAUDE.md citation
100%
100%
Rule text referenced
100%
100%
Multi-agent orchestration workflow
Lightweight eligibility check
37%
100%
Early exit conditions
0%
100%
Claude PRs not excluded
0%
100%
Lightweight CLAUDE.md discovery
0%
100%
Mid-tier PR summary agent
0%
100%
Four parallel review agents
0%
100%
Sonnet for CLAUDE.md compliance
0%
100%
Opus for bug detection
0%
100%
Validation step for bugs
50%
100%
Opus validates bugs, sonnet validates CLAUDE.md
37%
62%
PR context passed to subagents
25%
100%
CLAUDE.md path scoping
87%
100%
ea296ba
Table of Contents
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.