autopilot

Name: autopilot
Availability: InStock
Author: honigbart

Automate end-to-end plan execution and review across all phases.

1|Updated Apr 6, 2026

Search Tags:#automation #codex #ollama #plan-execution #autopilot #autonomous-review

Ready to run

npx skills add https://github.com/Honigbart/honeyflow --skill autopilot-honigbart

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: autopilot
Source: https://github.com/Honigbart/honeyflow/tree/main/autopilot
Command: npx skills add https://github.com/Honigbart/honeyflow --skill autopilot-honigbart

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates the end-to-end execution and review of active plans, enabling hands-free progress from the current phase to completion.

Core Features & Use Cases

autonomous execution and review across all plan phases, including integration with Claude/Codex reviews and optional Ollama input when available
maintains durable state across sessions using .ai/plans/<slug>/execution_state.md and .ai/plans/<slug>/session_log.md, and archives completed work to .ai/archive
supports the full autopilot loop with phase-by-phase execution, automated decision logging, and plan finalization

Quick Start

Start autopilot on an active plan to run all phases with autonomous execution and review until completion.

Frequently Asked Questions about autopilot

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate end-to-end plan execution hands-free?▼

Autonomous plan execution processes each phase sequentially using Codex, Claude, and optional Ollama for decision-making. It tracks progress via execution_state.md and session_log.md to ensure recoverability across sessions until finalization.

What is the process for autonomous plan review during execution?▼

Autonomous review uses Codex and Claude with optional Ollama input to evaluate decisions during execution. Automated decision logging and durable state tracking via session_log.md ensure full recoverability across all plan phases.

Do I need a final_plan.md file to start autonomous execution?▼

Yes, autonomous execution requires a final_plan.md file within an active plan directory under .ai/plans. The system uses this file alongside execution_state.md and session_log.md to process phases and finalize completed work.

How does autonomous execution handle state recovery across sessions?▼

State recovery relies on durable files like execution_state.md and session_log.md stored under .ai/plans. This architecture tracks progress and allows the autopilot execution loop to resume operations seamlessly without losing context.

Can I use Ollama for autonomous decision-making during plan execution?▼

Yes, Ollama provides optional input for autonomous decision-making during plan execution. The system integrates Ollama alongside Codex and Claude to evaluate phases, log decisions, and finalize completed work hands-free.

What are the limitations of hands-free plan execution automation?▼

Hands-free plan execution automation requires active plans with a final_plan.md file under .ai/plans. It cannot process plans lacking these structural files or operate outside the defined execution_state.md and archive tracking processes.