Runs one autonomous autopilot iteration: checks status, executes the predicted task, logs it, and schedules the next wake.
- What it does
- This skill performs a single autopilot loop iteration. It checks autopilot status, stops and disables autopilot when all tasks are complete or the iteration limit is reached, requests a prediction for the optimal next action, executes that task, logs the result, and schedules the next wake-up. It also documents cache-aware scheduling with a 270-second delay and lists task sources such as team tasks, swarm tasks, and markdown checklists.
- When to use it
- Use it when an autopilot loop needs to advance by one iteration, for example to pick up the next task from configured task sources and keep the loop running. It fits recurring autonomous progress checks where each run ends by scheduling the following wake-up.
- Requirements
- It relies on Claude Code's native /loop and the ScheduleWakeup tool, plus MCP tools for autopilot status, prediction, logging, progress, disable, and config. It ships no scripts; it is instructions only.
Run one autopilot iteration using Claude Code's native /loop:
- Check status:
mcp__plugin_ruflo-core_ruflo__autopilot_status
- If all tasks complete or max iterations reached, call
mcp__plugin_ruflo-core_ruflo__autopilot_disable and stop
- Get prediction:
mcp__plugin_ruflo-core_ruflo__autopilot_predict for the optimal next action
- Execute the predicted task (spawn agent, edit code, run tests, etc.)
- Log via
mcp__plugin_ruflo-core_ruflo__autopilot_log
- Schedule next:
ScheduleWakeup({ delaySeconds: 270, reason: "next autopilot iteration" })
Cache-Aware Scheduling
Always use delay 270s (under 300s cache TTL) to keep the prompt cache warm between iterations.
Task Sources
Autopilot discovers tasks from:
- team-tasks: Claude Code TaskList entries
- swarm-tasks: MCP task_list entries
- file-checklist: Markdown checkbox items in tracked files
Configure: mcp__plugin_ruflo-core_ruflo__autopilot_config({ taskSources: ["team-tasks", "swarm-tasks"] })
