Digs deep into code, elevates those who write it.
terminal
$ curl -fsSL https://getkaze.dev/mole/install.sh | bash $ mole init Digs deep into code, elevates those who write it.
$ curl -fsSL https://getkaze.dev/mole/install.sh | bash $ mole init 
Mole uses Claude to analyze every pull request, from style nits to architectural concerns. It classifies issues into 6 categories, scores code quality, and catches security vulnerabilities before they ship.
Code review shouldn't just find bugs. It should help developers grow. Mole tracks individual metrics, gamifies improvement with achievements and streaks, and gives tech leads a dashboard to see how the team evolves.
Automated PR reviews with inline comments and severity levels. Six categories: correctness, security, performance, maintainability, style, documentation.
Multi-pass analysis with Claude Opus that catches architectural issues and subtle bugs a single pass misses.
Automated scoring per PR and per developer. Track trends, set thresholds, detect vulnerabilities and hardcoded secrets.
Achievements, streaks, and leaderboards. Three personality modes: strict, balanced, encouraging.
Single binary backend. Compiles to ~15MB with zero runtime dependencies.
Reviews, scores, developer metrics, and team data. Standard schema, easy to backup.
Caching, rate limiting, and background job queues. Drop-in Redis replacement.
Haiku for fast reviews, Opus for deep analysis. Bring your own Anthropic API key.