etanhey/fleet-wrap
skills/golem-powers/fleet-wrap/SKILL.md
Quiet-down protocol for sprint close: when the fleet wraps, delete ALL polling crons and monitors, send ONE final dashboard + ONE message, then go SILENT. Use when: fleet wraps, all workers done, overnight queue exhausted, sprint close, Etan asleep/away with nothing approved left. Triggers: fleet wrap, wrap the fleet, stand down, going quiet, sprint close. NOT for: mid-sprint monitoring (keep your loops), spawning a successor (use /session-handoff first).
$ install --global
skillsauthnpx skillsauth add etanhey/golems fleet-wrapInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
Security Scan Results
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
3
Scanners Passed
9
Scanners in report
Clean
TrivyContainer and dependency vulnerability scanner
95%
Clean
SemgrepStatic code analysis for vulnerabilities
95%
Clean
mcp-scan (Snyk)Model Context Protocol security validation
95%
Skipped
Snyk (dep)Open source security scanning
50%
Skipped
Socket.devSupply chain security analysis
50%
Skipped
VirusTotalMulti-engine malware detection
50%
Skipped
CrowdStrikeAdvanced threat intelligence
50%
Skipped
OSV-ScannerOpen Source Vulnerability database check
50%
Skipped
OWASP Dep-Check
50%
Last scanned: Jun 7, 2026, 3:06 AM14.6s3 files scanned
SKILL.md
- name:
- fleet-wrap
- description:
- Quiet-down protocol for sprint close: when the fleet wraps, delete ALL polling crons and monitors, send ONE final dashboard + ONE message, then go SILENT. Use when: fleet wraps, all workers done, overnight queue exhausted, sprint close, Etan asleep/away with nothing approved left. Triggers: fleet wrap, wrap the fleet, stand down, going quiet, sprint close. NOT for: mid-sprint monitoring (keep your loops), spawning a successor (use /session-handoff first).
/fleet-wrap — the quiet-down protocol
Etan, at dawn after gen-10's overnight run (verbatim, red-team verified ✅RT): "Why were you just listing my messages in WhatsApp the whole night? Why didn't you stop?"
Recurring since gen-7 — idle-pane / wasted-money / heartbeat noise (
f3231646:[2655],62517efa:[7956]). gen-10 weave #24, imp10.
When the fleet wraps: ZERO polling crons, ONE final dashboard + ONE message, then SILENT.
If the planned overnight mechanism breaks, choose one of two paths only:
delegate real approved work to a live owner, or stand down SILENT. A polling loop
with zero workers is the inverted fleet-wrap failure: it burns the night reading
status/messages while nothing is actually being driven. Evidence: Etan at dawn,
verbatim raw operator turn: "Why were you just listing my messages in WhatsApp
the whole night? Why didn't you stop?" (orchestrator 62517efa:[7957]).
When this fires
- All workers are done/idle AND their claims are verified — no queued approved work remains.
- The overnight/sprint queue is exhausted.
- Etan said "wrap the fleet", "stand down", "we're done", or approved the final state.
Disambiguation vs /session-handoff: "wrap up" aimed at YOUR session (context
high, session ending, successor needed) routes to /session-handoff — that skill
owns single-session continuation. THIS skill fires on FLEET-wide close: all
workers done, nothing queued, the channel goes quiet. When both apply, run
/session-handoff FIRST — then this protocol governs the outgoing agent's exit.
Retire protocol (outgoing agent — pairs with /session-handoff)
Succession is routine (gen-12 evening: orc → gen-13 path, vl LEAD v1→v2, bl LEAD outgoing→successor — three successions one day). Before you compact or go silent:
- Hand off completely — live-watch transfer list written (see
/session-handoff); successor brief posted; grill answered (monitors still running during transfer). - Stop monitors — kill every cron/monitor you own so compaction doesn't fire on stale loops (
71a8e3f5 [5418]). - Final dashboard (if fleet-wide wrap via this skill).
- Kill your own pane — do not linger as a zombie orc writing duplicate handoffs.
Order: handoff → stop monitors → final dashboard (if fleet-wide wrap) → close pane.
The protocol (in order)
- VERIFY done. Per worker: state check + independent claim verification
(
/never-fabricateR8). A fleet is not wrapped because it looks quiet. - KILL ALL POLLING.
CronList→CronDeleteevery monitor/heartbeat/status cron you own;TaskStopbackground monitors; report the stopped IDs in the wrap-up (orc REF9). Zero exceptions — a "harmless" 5-minute status cron is exactly how the WhatsApp-all-night failure happened. - ONE final dashboard (
/html-dashboard, cloned template — never from scratch): sprint state, merged PRs, open loops, blockers, money/tokens spent. - ONE final message (WhatsApp / the agreed channel): dashboard link + a 2-3 line summary. ONE message. Not one per event.
- brain_store the wrap — sprint summary, stopped cron/monitor IDs, open
loops. Tags:
milestone,<project>. - SILENT. No further messages, no re-armed loops, no "still quiet" heartbeats — until Etan speaks or new approved work arrives.
What silence does NOT mean
- Blockers found during wrap go ON the dashboard and IN the one message — not as a follow-up stream.
- A real incident (data loss in progress, security) justifies ONE alert. Everything else waits for the next session.
- Silence ≠ abandonment. Open loops are written down (dashboard + collab + brain_store) so the next session picks them up — the channel goes quiet, the state does not evaporate.
Anti-patterns
| Don't | Evidence |
|---|---|
| Relaying/listing the user's own messages back to them overnight | The dawn quote above — the incident this skill exists for |
| Planned overnight mechanism broke, so orc kept polling with zero workers | Inverted fleet-wrap: delegate real work or stand down silent |
| Leaving the fleet-monitor cron running after wrap | "Why didn't you stop?" — polling crons outlive their sprint |
| "One more status update" after the final message | One message means one |
| Keeping idle worker panes alive "just in case" | Wasted-money panes, gen-7 f3231646:[2655] |
| Wrapping without verifying worker claims | A "wrapped" fleet with an unverified worker is an unwrapped fleet |
Integration
| Skill | Relationship |
|---|---|
| /orc | Composition-map row "Fleet wraps / sprint close" routes here; REF9 (stop monitoring when done) is step 2 |
| /session-handoff | Successor spawn happens BEFORE wrap; the outgoing agent then wraps silently |
| /weave | The weave fires at convergence — a weave run ends with its own fleet-wrap |
| /html-dashboard | The one final artifact (step 3) |
| /never-fabricate | R8 claim verification gates step 1 |
Related Skills
etanhey/phoenix-human-view
tools
VerifiedTrustedCommunity
The human-eval UX contract for Phoenix views: turn-by-turn scrollable replay (not a scorecard), hide-but-copyable IDs, collapsed thinking, identity chips, tool filters, tiny frozen starter datasets, mark-wrong-in-thread, mobile-first. Use when: building or reviewing ANY Phoenix/eval view, annotation UI, session replay, or human-grading surface. Triggers: phoenix view, eval UI, annotation view, session replay, human eval UX, grading interface. NOT for: Phoenix data pipelines/ingest (capture scripts have their own specs).
3SKILL.mdUpdated Jun 7, 2026
etanhey/mac-systems
tools
VerifiedTrustedCommunity
macOS systems specialist — AppKit NSPanel architecture, launchd services, socket activation, MCP bridge resilience, syspolicyd, and high-frequency SwiftUI dashboards. Use when building menu-bar apps, LaunchAgents, debugging syspolicyd/Gatekeeper/TCC, resilient UDS/MCP bridges, or SwiftUI dashboards at 10Hz+.
3SKILL.mdUpdated Jun 7, 2026
etanhey/judge-fleet
development
VerifiedTrustedCommunity
Bulk LLM-judging protocol for fleet-dispatched verdict runs (KG cluster, eval harness). Use when: dispatching or running judge workers (J1/J2/RT), planning bulk-apply from verdict JSONL, or triaging evidence_degraded outputs. Triggers: judge fleet, bulk judge, R3 verdicts, kg-judge, RT gate, evidence_degraded. NOT for: single-item code review, Phoenix view UX (use phoenix-human-view), or non-judge eval pipelines.
3SKILL.mdUpdated Jun 7, 2026
etanhey/drive-usage
development
VerifiedTrustedCommunity
Brain Drive filing discipline — where every artifact goes + how to name it. Use WHENEVER touching Google Drive / Brain Drive: uploading, creating folders, saving research prompts/results, audits, plans, transcripts, dashboards, or when about to leave a durable artifact in docs.local/. Teaches the numbered folder model (01_STANDARDS / 02_GROUNDING / 03_RESEARCH / 04_INGEST / 06_ARCHIVE), date-prefixed naming, and the rule: FILE durable artifacts in the right Drive folder — docs.local/ is cache-only. NOT for querying Drive via Gemini (use /braindrive) or web research (use /gemini-research); for >100KB heavy archival defer to /google-drive-archive.
3SKILL.mdUpdated Jun 7, 2026