pipeline

Use when designing before implementing — sketching types, signatures, and module structure for non-trivial work where jumping to code would lock in the wrong…

Use for each new Startup work event, including later requests, clarifications, multiple tasks and screenshot complaints; execution belongs to its assigned…

Use when a finished, reviewed change should be made as simple as it can be without changing behavior — a single behavior-preserving polish pass over a scoped…

Use when an operator keeps correcting agents for the same mistakes in a repo — find each repeated mistake class and make it impossible, preferring…

Use when verifying a live deployment — loading every page, checking console/network errors, exercising read-only functionality against a deployed base URL

Use when you want a repeatable, intake-driven code quality audit of a codebase — UI bugs, code bugs, cleanup/slop, or full audit.

Use when fixing a bug in place on your current branch — stays on the branch you are on, confirms the issue with an opt-in live repro, checks whether the bug…

Use when reviewing recent code changes, a pull request, or a git diff for a ship verdict — adversarial review from five angles with a club fight producing…

Use when doing server maintenance, patching, rebooting, or debugging a remote host, making changes, fixing issues, or recording completed operations on a…

Use when you want a scoped, multi-agent architectural audit of a specific subsystem (e.g., the auth system, the billing pipeline, the ingestion path) in a…

Use when the user requests Startup staff or the bound manager needs additional role capacity.

Use for an assigned Startup queue item of any kind, with its canonical procedure, independent review and outcome evidence.