# Hatchet CLI

> Hatchet CLI skills for managing workflows, workers, and runs. Use this skill when triggering workflows, starting workers, debugging runs, replaying failed runs, running Hatchet locally in embedded mode, or setting up the Hatchet CLI and profiles.

- Skill: `hatchet-dev/hatchet-cli` (Agent Skill, multi-file: 9 files)
- Install (CLI): `npx skillmds@latest add hatchet-dev/hatchet-cli`
- Raw SKILL.md: https://api.skillmd.com/api/skills/hatchet-dev/hatchet-cli/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Coding & Dev Tools
- License: MIT
- Author: hatchet-dev (https://skillmd.com/u/hatchet-dev)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/hatchet-dev/hatchet-cli

---


# Hatchet CLI Agent Skills

This skill package teaches AI agents how to use the Hatchet CLI to manage workflows, workers, and runs.

## When to use these skills

Read the relevant reference document before performing any Hatchet CLI task:

- **Setting up the CLI or creating a profile** → `references/setup-cli.md`
- **Running Hatchet locally without a token (embedded mode)** → `references/local-dev-embedded.md`
- **Starting a worker** → `references/start-worker.md`
- **Triggering a workflow and waiting for results** → `references/trigger-and-watch.md`
- **Debugging a failed or stuck run** → `references/debug-run.md`
- **Replaying a run with the same or new input** → `references/replay-run.md`

## Key conventions

- For local development, prefer embedded mode: no token, account, or server needed (`references/local-dev-embedded.md`).
- Always specify a profile with `-p HATCHET_PROFILE` unless a default profile is set. Embedded mode needs no profile.
- Use `-o json` for machine-readable output when parsing responses.
- Write workflow input to a temp file (e.g. `/tmp/hatchet-input-$(date +%s)-$$.json`) to avoid collisions.
- Clean up temp files after use.