squads pause/resume: first-class squad activation state, enforced by runner + org planner

[kokevidaurre]

Context (founder decision 2026-06-12)

Pause state already exists as SQUAD.md frontmatter (status: paused, paused_since, paused_reason — e.g. outreach since 2026-05-15) but is declarative only: nothing reads it. Known waste bug: the org planner dispatches paused squads anyway (fleet audit finding). Freezes currently work by agents happening to read prose in goals.md.

Spec

1. squads pause <squad> [--reason "..."] / squads resume <squad> — write the existing SQUAD.md frontmatter fields (no new state store).
2. Enforcement:
   • squads run <paused-squad> refuses with the reason + --force override for founder/manual use.
   • Org runner / planner SKIPS paused squads (fixes the waste bug; compose with the quota-aware runner from feat(org-run): quota-aware org runner — pre-flight probe, fail-fast, --wait-for-quota #861).
   • Cron/scheduled dispatch skips them too.
3. Visibility: squads status and squads dash show paused state + since/reason; squads status <squad> shows full pause metadata.
4. Docs/README regen after (same pipeline as context loader: single strategy file — rename directives slot, drop priorities slot, refresh README+docs #876).

Acceptance

• Pausing a squad makes it un-dispatchable by every automated path in one command, visibly, reversibly.

Related: #861 (org runner), #876 (context loader), hq governance epic #764 (advisory→mechanical pattern).

🤖 Generated with Claude Code