RUNBOOK BUILDER
Define the job.
Assemble the instructions. Review the output. Run it in your agent.
01
Outcome / Objective
Start with what should exist when the work is done.
02
Context
Give the agent the facts it needs.
03
Job / Scope
Define the work. One task or phase per line.
04
Constraints / Preserve
Protect the things that already work.
05
Execution Strategy
Choose how the work should proceed.
06
Autonomy
Allow routine decisions within the defined scope.
07
Approval Gates
Require explicit approval before consequential actions.
08
Return Conditions
Define when the agent should hand control back.
09
Failure / Recovery
Preserve progress when something goes wrong.
10
Definition of Done
Make completion objectively checkable.
11
Resource Strategy
Behavioral guidance, not a prediction of usage.
Your inputs stay in this browser. No account, AI API call, or cloud synchronization. Avoid shared devices for sensitive project context.