Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Gracefully shut down the AgentFlow pipeline. In plugin mode, also cleans up agent teams and uses SendMessage for faster shutdown signaling.
.claude/skills/bilal140202-sdlc-stop/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-02 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-04 | ✗→✓ | ▲ Improved | 16% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-07 | ✗→✓ | ▲ Improved | -24% | 0% |
| case-08 | ✗→✓ | ▲ Improved | 23% | 0% |
Gracefully shut down the AgentFlow pipeline.
In plugin mode, also:
Find the pinned Status task across all [SDLC] projects.
Post comment: [SYSTEM:PAUSING] Graceful shutdown initiated. Active workers will finish their current stage. No new tasks will be dispatched.
Check all tasks in Research/Build/Review/Test/Integrate stages:
For tasks with [BUILD:STARTED] and recent heartbeat (< 10 min):
For tasks in Research or Build WITHOUT [BUILD:STARTED] (not started yet):
[SLOT:T<N>] → [SLOT:--][STAGE:X] → [STAGE:Backlog]git worktree remove feat/<task-code>-<slug> --forceFor tasks in Review/Test that are waiting (no worker actively processing):
git worktree remove feat/<task-code>-<slug> --forceReport which tasks are still being worked on.
Tell the user: "Active workers are finishing. The system will be fully paused when all active work completes. You can check your Kanban board for current status."
Update the Status task:
[SYSTEM:PAUSED]
Graceful shutdown complete — <timestamp>
Tasks returned to Backlog: <N>
Tasks still completing: <N> (if any workers were mid-build)
To resume: run /sdlc-orchestrateTell the user:
To fully stop the orchestrator, comment out or remove the crontab entry:
crontab -e
# Comment out: */15 * * * * /usr/local/bin/claude -p "Run /sdlc-orchestrate" ...
To resume later, uncomment it and run /sdlc-orchestrate once to kick things off.[SYSTEM:PAUSING] status on their next board read and should stop picking up new work/sdlc-stop is idempotent — running it twice is harmless| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 5,850 | 6,141 | +5% | 1 | 1 | 0% | 269 | 1,193 | +343% | 0 | 0 | — |
case-02 | fail→pass | 13,014 | 10,890 | -16% | 1 | 1 | 0% | 2,535 | 2,941 | +16% | 0 | 0 | — |
case-03 | fail→fail | 10,904 | 5,560 | -49% | 1 | 1 | 0% | 2,308 | 989 | -57% | 0 | 0 | — |
case-04 | fail→pass | 7,641 | 3,799 | -50% | 1 | 1 | 0% | 1,228 | 1,428 | +16% | 0 | 0 | — |
case-05 | pass→pass | 71,227 | 4,383 | -94% | 1 | 1 | 0% | 1,923 | 1,569 | -18% | 0 | 0 | — |
case-06 | fail→pass | 6,011 | 2,750 | -54% | 1 | 1 | 0% | 1,032 | 1,280 | +24% | 0 | 0 | — |
case-07 | fail→pass | 11,250 | 3,770 | -66% | 1 | 1 | 0% | 1,990 | 1,512 | -24% | 0 | 0 | — |
case-08 | fail→pass | 5,955 | 2,878 | -52% | 1 | 1 | 0% | 1,052 | 1,292 | +23% | 0 | 0 | — |
case-09 | fail→pass | 8,657 | 3,113 | -64% | 1 | 1 | 0% | 1,544 | 1,335 | -14% | 0 | 0 | — |
case-14 | pass→pass | 14,226 | 6,607 | -54% | 1 | 1 | 0% | 2,316 | 1,845 | -20% | 0 | 0 | — |
case-10 | fail→pass | 8,980 | 4,677 | -48% | 1 | 1 | 0% | 1,833 | 1,658 | -10% | 0 | 0 | — |
case-11 | fail→pass | 19,155 | 2,064 | -89% | 1 | 1 | 0% | 1,818 | 1,124 | -38% | 0 | 0 | — |
case-12 | fail→pass | 15,873 | 2,451 | -85% | 1 | 1 | 0% | 2,507 | 1,163 | -54% | 0 | 0 | — |
case-13 | fail→pass | 10,244 | 3,465 | -66% | 1 | 1 | 0% | 1,784 | 1,321 | -26% | 0 | 0 | — |
case-15 | pass→pass | 10,777 | 2,960 | -73% | 1 | 1 | 0% | 1,846 | 1,219 | -34% | 0 | 0 | — |
case-16 | pass→pass | 11,954 | 6,204 | -48% | 1 | 1 | 0% | 1,994 | 1,744 | -13% | 0 | 0 | — |
case-17 | fail→pass | 7,488 | 3,580 | -52% | 1 | 1 | 0% | 1,402 | 1,346 | -4% | 0 | 0 | — |
case-18 | fail→pass | 8,861 | 1,891 | -79% | 1 | 1 | 0% | 1,499 | 1,008 | -33% | 0 | 0 | — |
case-19 | fail→pass | 8,839 | 6,821 | -23% | 1 | 1 | 0% | 1,304 | 2,114 | +62% | 0 | 0 | — |
case-20 | fail→pass | 9,639 | 2,126 | -78% | 1 | 1 | 0% | 1,378 | 999 | -28% | 0 | 0 | — |
case-21 | fail→pass | 11,064 | 3,942 | -64% | 1 | 1 | 0% | 1,866 | 1,405 | -25% | 0 | 0 | — |
case-22 | pass→pass | 7,483 | 4,098 | -45% | 1 | 1 | 0% | 1,340 | 1,326 | -1% | 0 | 0 | — |
case-23 | pass→pass | 10,262 | 5,624 | -45% | 1 | 1 | 0% | 1,882 | 1,795 | -5% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 23 cases were attempted, and 21 counted toward the lift figure. The other 2 produced results that are not comparable between the two arms, so they are excluded from the headline rather than averaged into it. The headline lift of +65 percentage points is the difference between those two pass rates over the 21 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
Other measured skills in the registry, with their headline benchmark lift.