From great_cto
Operates a gated autopilot for AI-driven workflows: start runs, view inbox of pending human signatures, approve or reject cases, and inspect run traces. Also supports list and status subcommands.
How this command is triggered — by the user, by Claude, or both
Slash command
/great_cto:autopilot start <vertical> [--live] [--tenant X] | inbox [--tenant X] | runs [--status S] | show <runId> | approve <runId> --by "<name>" | reject <runId> --by "<name>"haikuThis command is limited to the following tools:
The summary Claude sees in its command listing — used to decide when to auto-load this command
<!-- great_cto-managed --> You are the great_cto **/autopilot** command — the **Operate** side of the product. > Build vs Operate: `/start` + `/<vertical>-review` **build** an autopilot (the gated pipeline). > `/autopilot` **operates** one — runs a vertical's flow to its human checkpoint, holds the case in > an inbox for the licensed human (coder · BSA officer · broker · CPA · QPPV …), and executes the > irreversible action *only after* they sign. Inspect a flow first with `/flow <vertical>`. ## Dispatch ## Subcommands - `/autopilot start <vertical> [--live] [--tenant X]` — run the f...
You are the great_cto /autopilot command — the Operate side of the product.
Build vs Operate:
/start+/<vertical>-reviewbuild an autopilot (the gated pipeline)./autopilotoperates one — runs a vertical's flow to its human checkpoint, holds the case in an inbox for the licensed human (coder · BSA officer · broker · CPA · QPPV …), and executes the irreversible action only after they sign. Inspect a flow first with/flow <vertical>.
PD=$(ls -d ~/.claude/plugins/cache/local/great_cto/*/ 2>/dev/null | sort -V | tail -1 | sed 's|/$||'); [ -d "$PD/scripts" ] || PD=.
# Shared store with the admin board (~/.great_cto/autopilot-runs); override with GREAT_CTO_RUNS_DIR.
eval "node \"$PD/scripts/autopilot.mjs\" ${ARGUMENTS:-inbox}"
/autopilot start <vertical> [--live] [--tenant X] — run the flow to the first human gate; the
case enters the inbox awaiting-approval. --live exercises the real connectors./autopilot inbox [--tenant X] — the cases awaiting a signature (default action)./autopilot runs [--status awaiting-approval|completed|rejected] — all runs, newest first./autopilot show <runId> — the full step trace + connector results + gate status./autopilot approve <runId> --by "<name>" — sign the gate; the run resumes and the irreversible
action fires (multi-gate flows pause again at the next signature)./autopilot reject <runId> --by "<name>" — end the run; nothing irreversible runs.great-cto board → Autopilot) — the
CLI and the board are one queue, so a human can work from either.npx claudepluginhub avelikiy/great_cto/human-oversight-protocolApplies human oversight protocol to <scope>, enforcing approval gates, intervention commands, transparency requirements, action classification, and plan reviews.
/cc-autonomyDeploys autonomy operating profiles composing memory, workflow, orchestration, hooks, and verification into repo files and hooks. Supports deploy, status, switch, verify, plan subcommands with conservative to unattended-review profiles.
/fire-autonomousRuns the full build pipeline autonomously — plans, executes, verifies, and advances through all phases without human checkpoints after PRD is complete.
/loopExecutes a HOTL workflow with loop execution and auto-approval by invoking the hotl:loop-execution skill exactly as defined.
/automateCreates event-driven automations with triggers and actions using omni CLI. Tests with dry-run, lists automations, and views logs.