What happens if a verification step fails?

It enters a failure handling loop to diagnose and fix the error. If it cannot resolve the issue after a set number of attempts, it escalates to the user or uses specialized debugging skills.

How does this skill handle session loss or crashes?

The skill persistently tracks progress by updating checkboxes on GitHub issue comments and cross-referencing git logs, allowing it to resume from the exact point it left off.

Does it support human-in-the-loop workflows?

Yes, 'Guided Mode' provides regular checkpoints and pauses for user feedback, whereas 'Solo Mode' works autonomously to solve problems.

What are the requirements for verification?

Before completing a plan, the skill runs the project's full verification suite, including linting, type checking, and all unit/integration tests.

Does this skill perform automated commits?

Yes, it follows a discipline of committing each step individually with descriptive messages only after all verification checks for that step have passed.

Systematic Plan Execution

Name: Systematic Plan Execution
Author: thermiteau

bythermiteau

0•

생산성 및 워크플로

Executes multi-step implementation plans with persistent progress tracking, automated verification, and robust crash recovery.

The Maverick Plan Execution skill provides a disciplined framework for Claude to work through complex implementation plans. It manages the entire lifecycle of a code change—from picking the next task and implementing logic to running verification tests and committing work—while ensuring progress is saved via GitHub issue updates to survive session loss. By adapting between solo autonomous modes and guided human-checkpoint modes, it maintains codebase integrity through strict verification rules and systematic failure handling procedures.

주요 기능

01Integrated systematic debugging and escalation for failed verification steps

02Smart crash recovery and session resumption from the last verified step

03Automated execution loop covering implementation, verification, and commits

040 GitHub stars

05Persistent progress tracking using GitHub issue checkbox synchronization

06Dual execution modes for both autonomous and human-guided workflows

사용 사례

01Ensuring every step of a technical debt cleanup is verified and committed individually

02Executing complex multi-file feature migrations that require multiple sessions

03Automating the reporting of progress back to GitHub issues for team visibility

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add thermiteau/maverick-private mav-plan-execution

For use in Claude.ai and ChatGPT

주요 기능

01Integrated systematic debugging and escalation for failed verification steps

02Smart crash recovery and session resumption from the last verified step

03Automated execution loop covering implementation, verification, and commits

040 GitHub stars

05Persistent progress tracking using GitHub issue checkbox synchronization

06Dual execution modes for both autonomous and human-guided workflows

사용 사례

01Ensuring every step of a technical debt cleanup is verified and committed individually

02Executing complex multi-file feature migrations that require multiple sessions

03Automating the reporting of progress back to GitHub issues for team visibility

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add thermiteau/maverick-private mav-plan-execution

For use in Claude.ai and ChatGPT