From spacelift
This skill should be used when the user wants to manage Spacelift runs: confirm a pending run, discard a run, trigger a new deployment, retry a failed run, or cancel a queued run. Activates on: "confirm run", "confirm spacelift", "discard run", "discard spacelift", "trigger run", "trigger spacelift", "deploy spacelift", "retry spacelift", "cancel run", "cancel spacelift", "run spacelift", "lancia spacelift", "conferma run", "scarta run", "annulla run spacelift", "rilancia spacelift".
npx claudepluginhub fabn/claude-plugins --plugin spaceliftThis skill uses the workspace's default tool permissions.
Manage Spacelift runs: confirm, discard, trigger, retry, and cancel runs.
Creates isolated Git worktrees for feature branches with prioritized directory selection, gitignore safety checks, auto project setup for Node/Python/Rust/Go, and baseline verification.
Executes implementation plans in current session by dispatching fresh subagents per independent task, with two-stage reviews: spec compliance then code quality.
Dispatches parallel agents to independently tackle 2+ tasks like separate test failures or subsystems without shared state or dependencies.
Manage Spacelift runs: confirm, discard, trigger, retry, and cancel runs.
Reference files: Consult reference/spacectl-tools.md for the full MCP tool catalog and CLI equivalents.
spacelift.spacectl-mcp): confirm_stack_run, discard_stack_run, trigger_stack_run, list_stack_runs, list_stack_proposed_runs, get_stack_runspacectl (also covers retry and cancel not available via MCP)Use ToolSearch with query spacelift to check if MCP tools are available. See reference/spacectl-tools.md § Tool Detection.
| User intent | MCP Tool | CLI Equivalent |
|---|---|---|
| Confirm a pending run | confirm_stack_run | spacectl stack confirm --id <slug> --run <id> |
| Discard a pending run | discard_stack_run | spacectl stack discard --id <slug> --run <id> |
| Trigger a tracked deployment | trigger_stack_run (type: TRACKED) | spacectl stack deploy --id <slug> |
| Trigger a proposed run | trigger_stack_run (type: PROPOSED) | N/A |
| Retry a failed run | N/A (CLI only) | spacectl stack retry --id <slug> --run <id> |
| Cancel a queued run | N/A (CLI only) | spacectl stack cancel --id <slug> --run <id> |
If the user only provided a stack name:
list_stack_runs to find tracked runs pending confirmation (state UNCONFIRMED)list_stack_proposed_runs for preview runsAlways ask for confirmation before executing destructive actions:
<id> on <stack>?"<stack>?"Use get_stack_run and get_stack_run_changes to gather the details needed for the confirmation prompt.
Execute the action and report the result:
Run `<run-id>` on stack `<stack-slug>` confirmed. The apply phase is now in progress.
For triggered runs, report the new run ID so the user can track it.
| Situation | Action |
|---|---|
| MCP tools not available | Fall back to CLI. See reference/spacectl-tools.md § CLI Equivalents |
| Run not in expected state | Show current state — e.g., "Run is FINISHED, nothing to confirm" |
| No pending runs found | List recent runs and show their states |
| Permission denied | Check Spacelift account permissions and API key scopes |
| Authentication failure | Check SPACELIFT_API_GITHUB_TOKEN and SPACELIFT_API_KEY_ENDPOINT env vars |
reference/spacectl-tools.md — Full MCP tool catalog, CLI equivalents, GraphQL queries