Slash Commands

All available Candid commands.

/candid-review

Run a code review on your changes.

/candid-review

Flag	Description
`--harsh`	Use brutal honesty tone
`--constructive`	Use caring directness tone
`--focus <mode>`	Focus on security, performance, architecture, or edge-case
`--auto-commit`	Auto-commit applied fixes
`--re-review`	Compare against previous review
`--merge-target <branch>`	Specify the branch to diff against
`--exclude <pattern>`	Exclude files matching pattern

/candid-review --harsh --focus security
/candid-review --constructive --auto-commit
/candid-review --re-review

Run code review in a loop until all issues are resolved.

/candid-loop

Flag	Description
`--mode <auto\|review-each\|interactive>`	Execution mode (default: auto)
`--max-iterations <N>`	Maximum loop iterations (default: 5)
`--categories <list>`	Categories to enforce: critical, major, standards, smell, edge_case, architectural, or all (default: all)

/candid-loop
/candid-loop --mode review-each
/candid-loop --mode interactive
/candid-loop --max-iterations 3
/candid-loop --categories critical,major

Generate a Technical.md file by analyzing your codebase.

/candid-init

Flag	Description
`--template <name>`	Use a specific template (minimal, react, node, python)
`--effort <level>`	Analysis depth: quick (~30s), medium (~1-2min), thorough (~5-10min, default)
`--output <path>`	Custom output path (default: .candid/Technical.md)

quick: Fast analysis - framework detection, directory structure, file suffixes
medium: Balanced - adds naming conventions, error patterns, reads 5-8 key files
thorough: Comprehensive - launches sub-agents to read ALL files in parallel (default)

/candid-init
/candid-init --template react
/candid-init --effort quick
/candid-init --effort thorough

Check your Technical.md for issues.

/candid-validate-standards

Flag	Description
`--fix`	Auto-fix issues where possible