<CREW_NAME> standing rules

The controller's own file. Scope instructions: every thread in the crew reads it. Each rule that came from the owner keeps the date and the exact words, so lanes can cite it as a standing approval instead of asking again. Append; never rewrite an owner's words.

Mandate (<owner>, <date>)

"<the owner's words for what this crew is for>"

• Every <code prefix>-* thread in <project> is in the crew. New threads the owner starts are added too.
• Watch <REPO> issues and PRs. Start lanes for severe user-reported bugs no live thread has claimed (the severity bar is below).
• Severity first: S1 is a broken install or upgrade, data loss, a security result that is wrong, or BASE red. S2 is a broken workflow with no workaround. Everything else is batched.

Merging (<owner>, <date>: "<words>")

• One controller merges, through a train (/captain, the merge-train section). Lanes send READY and never merge. Standing approval to merge when green: "<words>".
• A merge outside a train costs a backstop gate; none without the owner's words.

Spend (<owner>, <date>: "<words>")

• Pre-approved: train gates, bisects, the backstop, screens, focused paid cases. Cap per rerun: <amount>. Over the cap is a fresh ask.

Asks to the owner (<owner>, <date>: "<words>")

• The ask format (gstack docs/captain.md), explained like I'm five: <Captain name>, then one lettered line per decision: "A: <code> hit problem p, which happens when h. We could fix it in these ways: 1) x; 2) y; 3) z. Recommendation: x because why. A1 to take recommendation." "do it" takes every recommendation in the list at full scope, follow-ons included, no further ask; a named option ("A2") takes that option instead.

Parallelism (<owner>, <date>: "<words>")

• Keep about <N> lanes working on disjoint files whenever there is backlog. Serial steps stay serial only where the tooling requires it: one train's slots, version stamps, edits to a shared file.

Decisions log

• <date> <time> · <decision> · "<owner's exact words>" · covers: <scope>