Quickstart
Run judge on a workspace, read the grouped report, drill into details and save a first baseline.
Run every command from the workspace root.
Run the combined report
cargo judge
The bare command runs every detector that needs no opt-in configuration. The terminal view is deliberately compact: findings are grouped by rule, split into evidence-backed findings and advisory heuristics (no verdict or score effect), each with a representative location, and followed by the next commands to run. The showcase shows the report for judge’s own source tree.
Colour is used only on an interactive terminal and honours NO_COLOR; --color always or
--color never overrides that.
Drill down
cargo judge --details # every finding and location
cargo judge dupes # clone families, grouped and prioritised
cargo judge refactor # ranked, evidence-backed review queue
cargo judge health --score # complexity, slop signals and a 0–100 score
cargo judge explain-rule swallowed-result # what a rule checks and where it stops
Keep the results
cargo judge --format json # writes .judge/judge.json
cargo judge --format markdown # writes .judge/judge.md for a pull request
cargo judge --save-baseline # writes .judge/baseline.json
Consider adding .judge/ to .gitignore unless you want to commit a baseline.
Compare later
cargo judge compare .judge/baseline.json
compare reports unchanged, resolved and introduced findings and a verdict. See
Baselines and CI for exit codes and a CI setup.