throughput
Use when the user asks to open/show the CCC throughput (token) dashboard, or asks how many tokens they've used — overall/so far,…
Use after a UI or user-visible change and before you claim it works — it spawns ONE CCC verification lane that drives a real browser (gstack browse, or CCC's puppeteer snapshot as fallback) against the running app, checks the specific thing you changed, and reports a
$ npx -y skills add amirfish1/claude-command-center --skill fleet-verify --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/fleet-verifyContext preview
The summary Claude sees to decide when to auto-load this skill.
Use after a UI or user-visible change and before you claim it works — it spawns ONE CCC verification lane that drives a real browser (gstack browse, or CCC's puppeteer snapshot as fallback) against the running app, checks the specific thing you changed, and reports a
name: fleet-verify description: Use after a UI or user-visible change and before you claim it works — it spawns ONE CCC verification lane that drives a real browser (gstack browse, or CCC's puppeteer snapshot as fallback) against the running app, checks the specific thing you changed, and reports a visual/behavioral verdict with a screenshot. Independent of your context; catches what a diff read and a unit test cannot. allowed-tools: Bash
`/code-review` reads the diff. `pair-verify` runs a bug at two refs. Neither opens the app and *looks*. Fleet-verify spawns one CCC lane that drives a headless browser against the running dashboard (or your app), verifies the exact change you describe, captures a screenshot as evidence, and reports a verdict. It pairs CCC's spawn/report_to mechanics (`ccc-orchestration`) with gstack `browse` (headless Chromium QA).
**1 spawned session.** A real billed session on the kanban — state the count before spawning. Skip it for pure backend/logic changes with no rendered surface (there is nothing to look at); use a test or `pair-verify` instead.
The lane can only check what you can point it at. Have all three:
`http://127.0.0.1:8090`. For another app, its dev-server URL. If nothing is running, start it first (or say so in the prompt so the lane starts it).
status rail lists each subagent", not "check the UI". Name the selector, text, or state you expect.
CCC_URL="$(cat ~/.claude/command-center/port.txt 2>/dev/null || echo "${CCC_URL:-http://127.0.0.1:8090}")"
curl -sf --max-time 2 "$CCC_URL/api/version" >/dev/null 2>&1 || CCC_URL="http://127.0.0.1:8090"
curl -sf --max-time 2 "$CCC_URL/api/version" >/dev/null 2>&1 || CCC_DOWN=1in the Bash sandbox and fails spuriously even when CCC is up).
curl -s -X POST "$CCC_URL/api/sessions/spawn" -H "Content-Type: application/json" -d @- <<'JSON'
{
"prompt": "Independent visual verification. Drive a headless browser and confirm a specific change. 1) Use gstack browse (the `browse` skill): `$B goto <url>`, then `$B text`, `$B js \"<expr>\"`, `$B is visible @<ref>` to check state, and `$B screenshot /tmp/fleet-verify-<slug>.png` for evidence. If gstack browse is unavailable, run CCC's `node snapshot.js` first (it writes snapshot.png). Do not substitute chrome-devtools: only use chrome-devtools after `curl -fsS --max-time 2 http://127.0.0.1:<debug-port>/json/version` returns Chrome JSON. 2) Verify EXACTLY this: <the specific thing — selector/text/state you expect>. 3) Report VERDICT (VERIFIED / NOT-VISIBLE / WRONG-STATE / APP-DOWN), the commands you ran, and the screenshot path as evidence. Be literal: report what you actually see, not what the change intended.",
"repo_path": "/abs/path/to/repo",
"report_to": "<your-session-id>"
}
JSONyou pass either, pass `"engine"` too (both are validated per engine, and the effort ladder differs by engine: see `ccc-orchestration` § 2).
back to you when the lane finishes.
After spawning, **end your turn.** Tell the user a lane is verifying and give its session id. The verdict arrives by injection — do **not** poll or sleep-loop. If it never arrives, check `GET /api/sessions/spawned` for whether the lane is alive.
prove it. Ship it.
did not take, or you pointed it at the wrong place. Investigate before claiming.
If the arguments contain `dry-run`, print the exact payload — session count (1), the filled-in verifier prompt with URL and the specific check, and the target repo — and POST nothing.
yourself against the running app, and in your claim say verification was **self-run** — you know what the change was meant to do, so you are a weaker skeptic than a fresh lane. Never pretend the spawn ran.
check (curl the endpoint, assert the JSON/DOM string). A string assertion is not the same as *seeing* it render — flag the gap.
Your coding agents outgrew your terminal. CCC puts every session on one local board and tells you which one needs you. Not affiliated with Anthropic. Start the next while Claude builds the first. A real share card from one of the maintainer's machines.
Use when the user asks to open/show the CCC throughput (token) dashboard, or asks how many tokens they've used — overall/so far,…
Use when you want a marketing-copy A/B — spawns two writer sibling sessions via CCC that draft the SAME message architecture in…
Use when a bug has already survived at least one debugging attempt and you have 2-3 genuinely distinct root-cause hypotheses —…
Spawn, inject into, and ask questions of persistent sibling sessions via Claude Command Center (CCC).
Use to catch documentation that has silently gone stale — it spawns ONE fresh sibling session via CCC that extracts every…
Use for docs QA — it spawns ONE fresh sibling session via CCC that has never seen the feature, has it follow the README /…