warden run

warden [targets...] [options]
warden run [targets...] [options]

Runs Warden analysis against explicit targets or the current branch by default. The bare warden command is an alias for this command.

Arguments

Argument	Purpose
targets...	Files, globs, git refs, or @files containing target lists.

Options

Option	Purpose
-C, --cwd <path>	Run as if invoked from this directory.
--skill <name|path>	Run only this skill by name or path. Names fall back to built-ins.
--config-path <path>	Path to warden.toml, or a directory containing it. --config is a deprecated alias.
-m, --model <model>	Model fallback when warden.toml does not specify one. See Models and Runtimes.
--effort <level>	Override effort level for this run. Values: off, low, medium, high, xhigh, max.
--runtime <claude|pi>	Runtime backend for model-backed execution.
--json	Output results as JSON.
-o, --output <path>	Write full run output to a JSONL file.
--fail-on <severity>	Exit with code 1 if findings meet this severity.
--report-on <severity>	Only show findings at or above this severity.
--min-confidence <level>	Only show findings at or above this confidence.
--fix	Automatically apply all suggested fixes.
--parallel <n>	Max concurrent hunk analyses across running skills.
-x, --fail-fast	Stop after the first finding.
--staged	Analyze only staged changes.
--git	Force ambiguous targets to be treated as git refs.
--offline	Use cached remote skills and skip Pi model-catalog network refresh.
--quiet	Errors and final summary only.
-v, --verbose	Increase verbosity. Repeat as -vv or combine with --debug.
--debug	Enable debug output.
--log	Use log output mode.
--color / --no-color	Override color detection.
-h, --help	Show help for this command.

Examples

warden src/auth.ts
warden @targets.txt --skill security-review
warden "src/**/*.ts" --skill security-review
warden HEAD~3
warden --staged --fix