rundown

name	rundown
description	Use when asked to run or start a Rundown runbook by name (e.g. "run the planning runbook", "start the deploy runbook") and no runbook is active yet. The launcher that resolves a runbook and begins execution. Invocable as /rundown <runbook>.

Rundown Launcher

The entrypoint for starting a runbook from a cold start. This skill resolves a runbook by name, starts it, and hands off to the execution protocol. It does not contain the protocol itself.

When to Use

• A user asks to "run"/"start the X runbook" and no runbook is active yet.
• Invoked directly as /rundown <runbook> with a runbook name.

When NOT to Use

• A runbook is already active or CLI output asks for pass/fail — use running-runbooks.
• Authoring or editing a runbook file — use writing-runbooks.
• A runbook has its own bootstrap skill (e.g. planning) — invoke that skill; it starts itself.

Steps

1. Resolve the runbook. If the name is ambiguous or you are unsure it exists, list discoverable runbooks:

   rundown ls --all
   

   Names support namespace:name (e.g. rundown:planning for the plugin source). A bare name resolves via the priority chain (project → plugin → bundled).

2. Load the execution protocol — before starting. So you are ready to interpret the first step's output (including a delegation) the moment it appears, rather than scrambling for the protocol after rundown run has fired:

   Skill(skill: "rundown:running-runbooks")
   

   If the runbook contains a - DELEGATE step, also load the delegation protocol:

   Skill(skill: "rundown:delegating-runbooks")
   

3. Start it.

   rundown run <name>
   

   Capture the run id when you start the runbook: rundown run prints it at start and every subsequent event carries it as runbookId. You need it for every mutating command you issue as the orchestrator of a delegation-exposed run.

   Pass inputs if the runbook requires them:

   rundown run <name> --input key=value
   

   If rundown run reports missing required inputs, supply them and re-run.

4. Execute. Follow the loaded running-runbooks protocol: respond to each step with rundown pass / rundown fail and trust Rundown for transitions. On a delegation-exposed run (any runbook with a - DELEGATE step), these become rundown pass --claim-id <claim_id> / rundown fail --claim-id <claim_id> — the bare form is for standalone runs only and otherwise refuses with ACTOR_CONTEXT_REQUIRED.

Reference

• running-runbooks — the execution protocol
• writing-runbooks — authoring runbooks