Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Report AgentGuards guardrail status. Use when the user asks whether AgentGuards is active, healthy, or correctly configured in Codex.
.claude/skills/hashgraph-online-status/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-06 | ✗→✓ | ▲ Improved | 1190% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -20% | 0% |
| case-09 | ✗→✓ | ▲ Improved | -36% | 0% |
| case-10 | ✗→✓ | ▲ Improved | -37% | 0% |
| case-11 | ✗→✓ | ▲ Improved | -33% | 0% |
Display the current state of the team's work.
Use the --progress flag for a compact, human-friendly view:
bashpython3 <runtime>/agenteam_rt.py status --progress
If the user specifically asks for raw JSON state, use without --progress:
bashpython3 <runtime>/agenteam_rt.py status
Display the progress view in a readable format:
AgenTeam Run: <run-id>
Task: <task description>
Profile: <profile or "full">
Status: <running|completed|failed|stopped>
Elapsed: <Nm Ss>
Stages:
research ✓ completed (0m 45s)
strategy ✓ completed (0m 30s)
design ✓ completed (1m 02s)
implement → verifying (1m 15s) [verify attempt 2/3]
test · pending
review · pending
Active Lock: dev
Last Event: stage_verified (implement) — fail, attempt 2
Attempt: implement-dev-a1 | thread_id: <codex-thread-id>
Last heartbeat: 12s ago | idle budget: 14m 48s remaining
Wall budget: 47m remaining | stop reason: (none)The progress view includes elapsed times per stage, the current verify attempt if applicable, and the most recent event for context. When governance.adoption is present, include the high-signal governance items in the readable status: open follow-ups/escalations, recorded tripwire blocks or warnings, gate rejections, and criteria overrides. Do not ask the user to inspect raw .agenteam/governance files for these summary facts. For an active or interrupted attempt, always include thread_id (or "unavailable"), last heartbeat and age, idle budget and wall budget remaining, attempt/retry counts, PID liveness when known, and stop reason. Treat a stale heartbeat as AT RISK, not as proof that the role failed.
If no run is found, show:
.agenteam/config.yaml or legacy agenteam.yaml exist?)$ateam:run to start a new task."✓ — completed→ — in progress· — pending✗ — failed/blocked| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→fail | 4,212 | 16,062 | +281% | 1 | 1 | 0% | 694 | 844 | +22% | 0 | 0 | — |
case-02 | fail→fail | 14,328 | 11,988 | -16% | 1 | 1 | 0% | 1,216 | 860 | -29% | 0 | 0 | — |
case-03 | fail→fail | 10,272 | 16,320 | +59% | 1 | 1 | 0% | 1,184 | 855 | -28% | 0 | 0 | — |
case-04 | pass→fail | 13,654 | 19,990 | +46% | 1 | 1 | 0% | 1,146 | 852 | -26% | 0 | 0 | — |
case-05 | pass→pass | 5,898 | 11,038 | +87% | 1 | 1 | 0% | 1,042 | 1,577 | +51% | 0 | 0 | — |
case-06 | fail→pass | 5,686 | 18,018 | +217% | 1 | 1 | 0% | 229 | 2,954 | +1190% | 0 | 0 | — |
case-07 | pass→fail | 17,499 | 13,398 | -23% | 1 | 1 | 0% | 1,744 | 1,968 | +13% | 0 | 0 | — |
case-08 | fail→pass | 8,781 | 8,452 | -4% | 1 | 1 | 0% | 1,318 | 1,048 | -20% | 0 | 0 | — |
case-09 | fail→pass | 12,562 | 9,278 | -26% | 1 | 1 | 0% | 2,023 | 1,286 | -36% | 0 | 0 | — |
case-10 | fail→pass | 13,260 | 7,365 | -44% | 1 | 1 | 0% | 1,383 | 870 | -37% | 0 | 0 | — |
case-11 | fail→pass | 19,574 | 10,005 | -49% | 1 | 1 | 0% | 1,926 | 1,284 | -33% | 0 | 0 | — |
case-12 | pass→pass | 17,370 | 4,458 | -74% | 1 | 1 | 0% | 1,960 | 1,434 | -27% | 0 | 0 | — |
case-17 | fail→pass | 13,814 | 2,124 | -85% | 1 | 1 | 0% | 1,304 | 833 | -36% | 0 | 0 | — |
case-13 | pass→pass | 10,236 | 7,214 | -30% | 1 | 1 | 0% | 717 | 839 | +17% | 0 | 0 | — |
case-14 | fail→pass | 15,606 | 8,162 | -48% | 1 | 1 | 0% | 1,567 | 992 | -37% | 0 | 0 | — |
case-15 | fail→pass | 18,736 | 6,874 | -63% | 1 | 1 | 0% | 1,855 | 764 | -59% | 0 | 0 | — |
case-16 | fail→pass | 12,642 | 6,316 | -50% | 1 | 1 | 0% | 1,010 | 779 | -23% | 0 | 0 | — |
case-18 | fail→pass | 19,294 | 11,336 | -41% | 1 | 1 | 0% | 2,075 | 1,789 | -14% | 0 | 0 | — |
case-19 | pass→pass | 16,748 | 7,472 | -55% | 1 | 1 | 0% | 1,535 | 982 | -36% | 0 | 0 | — |
case-20 | fail→pass | 2,148 | 3,160 | +47% | 1 | 1 | 0% | 361 | 1,147 | +218% | 0 | 0 | — |
case-21 | pass→pass | 13,261 | 8,274 | -38% | 1 | 1 | 0% | 1,928 | 1,131 | -41% | 0 | 0 | — |
case-22 | pass→fail | 15,492 | 3,395 | -78% | 1 | 1 | 0% | 2,133 | 1,124 | -47% | 0 | 0 | — |
case-23 | pass→pass | 16,567 | 10,521 | -36% | 1 | 1 | 0% | 1,633 | 1,359 | -17% | 0 | 0 | — |
case-24 | pass→pass | 19,377 | 4,016 | -79% | 1 | 1 | 0% | 2,245 | 1,296 | -42% | 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. 24 cases were attempted, and 19 counted toward the lift figure. The other 5 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 +33 percentage points is the difference between those two pass rates over the 19 comparable cases. 3 cases got worse with the skill loaded, and they are included in that figure.
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.