How this command is triggered — by the user, by Claude, or both
Slash command
/wave:wave-helpThe summary Claude sees in its command listing — used to decide when to auto-load this command
# Wave Orchestration Help ## Commands | Command | Description | |---------|-------------| | `/wave "goal"` | Start wave orchestration with a goal | | `/wave-status` | Check current iteration and progress | | `/cancel-wave` | Stop the orchestration loop | | `/cancel-wave --clean` | Stop and clean up output files | | `/wave-preflight` | Check if environment is ready | | `/wave-clean` | Remove wave output files | | `/wave-help` | Show this help | ## Usage ### Options - `--max-iterations N` — Stop after N iterations (default: 20) - `--completion-promise TEXT` — Phrase that signals comple...
| Command | Description |
|---|---|
/wave "goal" | Start wave orchestration with a goal |
/wave-status | Check current iteration and progress |
/cancel-wave | Stop the orchestration loop |
/cancel-wave --clean | Stop and clean up output files |
/wave-preflight | Check if environment is ready |
/wave-clean | Remove wave output files |
/wave-help | Show this help |
/wave "Your goal here" [--max-iterations N] [--completion-promise TEXT] [--clean]
--max-iterations N — Stop after N iterations (default: 20)--completion-promise TEXT — Phrase that signals completion (default: DONE)--clean — Clear previous wave outputs before starting# Build a feature
/wave "Build user authentication with JWT, tests must pass" --max-iterations 15
# Research a codebase
/wave "Create architecture documentation for this codebase" --max-iterations 5
# With custom completion signal
/wave "Fix all TypeScript errors" --completion-promise "NO_ERRORS"
# Start fresh, clearing old outputs
/wave "Refactor database layer" --clean
/wave "goal"Output <wave-complete>PROMISE</wave-complete> when truly done.
Example:
All tests passing, feature complete.
<wave-complete>DONE</wave-complete>
.claude/wave-state.local.md — Iteration and config.claude/wave-outputs/ — Agent resultsBackground agents need tool permissions. If agents fail with permission errors:
claude --dangerously-skip-permissionsRun /wave-preflight to check your environment.
Wave won't start: Check if one is already running with /wave-status
Agents hitting permission errors: Run /wave-preflight and approve tools
Stuck in loop: Use /cancel-wave or wait for max iterations
Want to see progress: Use /wave-status for a visual progress bar
npx claudepluginhub sethdford/claude-toolkit --plugin wave3plugins reuse this command
First indexed Jan 9, 2026
/delegateOrchestrates complex tasks: decomposes into agent waves/tasks, plans execution, awaits user approval, then executes via subagents or teams.
/orchestrateRuns the orchestrator pattern on any codebase: fans out parallel subagents across waves of work, verifies between waves, and commits wave-by-wave. In Recon mode (no args), produces a prioritized findings list first.
/fast-forwardFast-forwards remaining nWave phases (DISCUSS, DESIGN, DEVOPS, DISTILL, DELIVER) end-to-end without review pauses after user confirmation. Optional feature description or --from flag.
/runParses and executes legacy inline orchestration workflow syntax supporting raw operators, YAML frontmatter templates, and temporary AI agents with model overrides.
/orchestrateRuns a sequential multi-agent workflow for complex tasks — planning, implementation, review, and security audit — with structured handoffs and a final report.
/workflowOrchestrates a full multi-agent development pipeline: design, planning, plan review, implementation, and code review. Produces tested, reviewed code with commit and pipeline metrics.