-
-
Notifications
You must be signed in to change notification settings - Fork 0
CLI Usage
🌍 English narrative is auto-generated on release. See the French version for the current full content.
WorkPilot AI ships a CLI for headless use and CI/CD integration.
Source: `docs/CLI-USAGE.md`
For headless operation, CI/CD pipelines, or terminal-only workflows, WorkPilot AI's full functionality is available without the desktop UI through the Python backend CLI.
CLI usage requires Python 3.9+ and the Claude Code CLI (pnpm install -g @anthropic-ai/claude-code). From apps/backend, create a virtual environment (uv venv && uv pip install -r requirements.txt, or the standard venv equivalent), then copy .env-files/.env.example to .env-files/.env and add an OAuth token obtained via claude setup-token.
| Step | Command |
|---|---|
| Create a spec |
python runners/spec_runner.py --interactive or --task "..."
|
| List specs | python run.py --list |
| Run a build | python run.py --spec 001 |
| QA validation | python run.py --spec 001 --qa |
| Review changes | python run.py --spec 001 --review |
| Merge to project | python run.py --spec 001 --merge |
| Discard build | python run.py --spec 001 --discard |
Spec creation automatically assesses complexity and assigns a tier — SIMPLE (3 phases, 1–2 files), STANDARD (6 phases, 3–10 files), or COMPLEX (8 phases, multiple services and integrations) — which determines how many pipeline phases run.
Every build runs in an isolated Git worktree under .worktrees/, so changes can be tested independently before merging into the main project. After QA validation completes (or is skipped with --skip-qa), the QA Reviewer and QA Fixer loop automatically until acceptance criteria pass, up to 50 iterations.
While a build is running, Ctrl+C once pauses it for added instructions, and pressing it twice exits. Alternatively, a PAUSE file dropped into the spec directory pauses after the current session, and instructions can be injected via a HUMAN_INPUT.md file.
Key environment variables include CLAUDE_CODE_OAUTH_TOKEN (required), AUTO_BUILD_MODEL, DEFAULT_BRANCH, plus optional integration keys (LINEAR_API_KEY, GITLAB_TOKEN) and Graphiti memory settings (GRAPHITI_ENABLED, provider keys). Specs can be validated independently with python validate_spec.py --spec-dir specs/001-feature --checkpoint all.
➡️ Continue: Security
Getting started / Pour débuter
- 🏠 Home
- 📘 Introduction · fr
- ⚡ Installation · fr
- 🚀 First project · fr
- 🧠 Key concepts · fr
- ❓ FAQ · fr
Usage
- 🖥 User interface · fr
- 📝 Creating a spec · fr
- 🔁 Multi-agent pipeline · fr
- 🤖 Specialized agents · fr
- 🔌 Integrations · fr
- 💡 AI providers · fr
- 🧩 Memory system · fr
Advanced / Avancé
Community / Communauté