Can I stop Fleet Co-Pilot once it has started working?

Yes, you can manually interrupt the process using the unlock command or if the agent triggers an escalation when it encounters a problem it cannot solve alone.

What makes Fleet Co-Pilot different from standard Claude?

Unlike standard interactions, Fleet Co-Pilot uses a 'lock mode' and SessionCopilot reasoning to work iteratively toward a specific objective without requiring a new prompt for every individual action.

Does Fleet Co-Pilot support automated testing?

Yes, you can include 'passing tests' as part of the Definition of Done, and the agent will use your local testing tools to verify its work before completing the task.

How does the agent know when a task is finished?

At the start of a session, the skill creates a 'Definition of Done' file. The agent continuously monitors its own progress and only stops once every criterion in that file has been verified.

Fleet Co-Pilot

Name: Fleet Co-Pilot
Author: rysweet

byrysweet

•

生産性とワークフロー

Automates complex coding tasks by formulating structured goals and working autonomously until specific completion criteria are met.

Fleet Co-Pilot transforms Claude into a goal-oriented autonomous agent that manages its own development cycle from initial instruction to final verification. By translating natural language prompts into a formal 'Definition of Done,' it enters a persistent reasoning mode that monitors progress across multiple steps without requiring constant user intervention. This skill is particularly effective for long-running refactoring tasks, multi-file bug fixes, or feature implementations where the agent needs to iterate independently until a verified outcome is achieved.

主な機能

01Automated reasoning and progress monitoring via SessionCopilot

02Persistent 'Lock Mode' for uninterrupted multi-step task execution

03Smart escalation triggers when human intervention is required

04Structured 'Definition of Done' criteria for verifiable task completion

05Autonomous goal formulation from natural language instructions

0645 GitHub stars

ユースケース

01Implementing new features from high-level descriptions with automated verification

02Resolving complex bugs that require iterative testing and correction across multiple files

03Executing large-scale codebase refactors that require consistent logic application

主な機能

01Automated reasoning and progress monitoring via SessionCopilot

02Persistent 'Lock Mode' for uninterrupted multi-step task execution

03Smart escalation triggers when human intervention is required

04Structured 'Definition of Done' criteria for verifiable task completion

05Autonomous goal formulation from natural language instructions

0645 GitHub stars

ユースケース

01Implementing new features from high-level descriptions with automated verification

02Resolving complex bugs that require iterative testing and correction across multiple files

03Executing large-scale codebase refactors that require consistent logic application