helix-e2e

# End-to-end test of a Helix deployment The "does the whole thing actually work" pass: org → project → agent → task → chat → approve → PR → cleanup. Run it after an install, after an upgrade, or when a bug report needs a clean reproduction. Prerequisites: a running control plane ([helix-deploy](../helix-deploy/SKILL.md)), a `helix` binary, and a user API key. ## Try the built-in harness first For a post-upgrade check that doesn't need a new org, the CLI ships its own harness — this is the cheap path, and it works on every version: ```bash helix spectask e2e --project "$PROJECT" --agent "$AGENT" \ --prompt "List the files in the repo" --cleanup helix spectask test --session "$SESSION" --all --json ``` `spectask e2e` creates a task, boots the sandbox, cancels a live turn, sends a follow-up, screenshots, and exercises the session MCP tools — a good regression check for the sandbox path specifically. It does **not** cover org creation, project YAML, spec approval or the PR. ## The full pass Copy this checklist and tick items off as you go: ``` - [ ] 0. Account and API key - [ ] 1. Organization - [ ] 2. Project (with a repository attached) - [ ] 3. Agent, and pre-flight the sandbox runner - [ ] 4. Dispatch a task - [ ] 5. Wait for the sandbox - [ ] 6. Chat with the running agent - [ ] 7. Board mechanics - [ ] 8. Approve and check the PR - [ ] 9. Tear down ``` Every step, with the commands and what each one proves: [reference/walkthrough.md](reference/walkthrough.md). When a step fails, match the symptom: [reference/failures.md](reference/failures.md). Two failures account for most first runs — the project has no repository attached (step 2), and there is no sandbox node registered (step 5). Check both before anything else: ```bash helix api /projects/$PROJECT/repositories # [] means step 2 will fail helix api /sandboxes # [] means step 5 will never finish ```