L’interface est localisée ; les présentations et documents des Skills sont traduits automatiquement par IA dans votre langue. Le code, les commandes, chemins et termes techniques restent en version originale, et le document de l’auteur fait foi.
Explorer les Skills
Déploiement et opérations · SKILL.md

trae-claw-install

trae-claw-install

Run a repository-driven OpenClaw deployment workflow with platform routing, setup/start/check, acceptance checks, and unified troubleshooting.

Independent GitHub snapshot · Automated classification · Original documentation governs

Documentation de la Skill

GitHub

name: trae-claw-install description: Run a repository-driven OpenClaw deployment workflow with platform routing, setup/start/check, acceptance checks, and unified troubleshooting.

TraeClawInstall

Description

This skill executes an end-to-end OpenClaw deployment workflow from a repository that already provides platform scripts and operational documents.

Usage Scenario

Use this skill when:

  • The user asks for one-stop OpenClaw local deployment.
  • The user wants to follow a repository-standard install/start/check flow.
  • The user has setup/start failures and needs structured troubleshooting.

Preconditions

  • Current working directory is the repository root.
  • Terminal execution is available.
  • Repository contains platform scripts and troubleshooting docs.

Instructions

  1. Detect platform and route scripts:
    • Windows: prefer WSL2 and use scripts/windows/wsl/*.sh
    • macOS: use scripts/macos/*.sh
    • Linux: use scripts/linux/*.sh
  2. Validate baseline:
    • node --version major version is >=22
    • npm --version is available
    • if openclaw --version is missing, continue with setup step
  3. Execute standard flow:
    • setup -> start -> check
  4. Run minimum acceptance:
    • openclaw doctor
    • openclaw status
    • openclaw dashboard
  5. If any step fails, run troubleshooting workflow:
    • capture the first error line
    • rerun doctor/status
    • verify binary paths and versions
    • apply fixes from repository troubleshooting docs
    • change one variable at a time and re-verify

Output Contract

  • Success: platform, executed steps, acceptance results, and service accessibility.
  • Failure: first error, diagnostics performed, and next actionable fix.

Constraints

  • Reuse repository scripts and docs; do not create a parallel flow.
  • Never write real secrets; only use example configuration.
  • On Windows, prefer execution inside WSL2 Linux filesystem.
Skills enregistrées