compose:execute

Executes written implementation plans with task tracking, verification steps, and review checkpoints.

Updated Jun 27, 2026
One-click install
npx skills add https://github.com/HKUST-QUANT-SOCIETY/quantcode --skill compose-execute-hkust-quant-society
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: compose:execute
Source: https://github.com/HKUST-QUANT-SOCIETY/quantcode/tree/main/.opencode/meta-skills/execute
Command: npx skills add https://github.com/HKUST-QUANT-SOCIETY/quantcode --skill compose-execute-hkust-quant-society

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Turning a written implementation plan into completed code often fails because steps get skipped, verifications are ignored, or blockers are silently worked around. This Skill enforces a disciplined execution loop: load the plan, review it critically, execute each task with its specified verifications, and stop to ask when blocked. ## Core Features & Use Cases - Critical Plan Review: Reads the plan file first and raises concerns with the user before writing any code. - Step-by-Step Task Execution: Marks tasks in progress, follows bite-sized steps exactly, runs specified verifications, and marks tasks completed. - Structured Blocker Handling: Stops immediately on missing dependencies, failing tests, or unclear instructions and presents options via compose:ask instead of guessing. - Use Case: After drafting a multi-step refactoring plan, hand it to this Skill to implement each task in an isolated worktree, verify every change, and produce a final report through compose:report before merging. ## Quick Start Execute the implementation plan in plan.md using the compose:execute skill, verifying each task and stopping to ask if anything is blocked.

Frequently Asked Questions about compose:execute

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

FAQPage Schema
How do I execute an implementation plan with an AI coding assistant?

Provide the plan file and invoke the compose:execute skill. It reviews the plan critically, raises concerns before starting, then executes each task in order, running the verifications specified in the plan and marking tasks complete as it goes.

What happens when plan execution hits a blocker or failing test?

Execution stops immediately when a dependency is missing, a test fails, or an instruction is unclear. The skill uses compose:ask to present the blocker and options rather than guessing, and only continues with best judgment if no user is available.

Should I use compose:execute or compose:subagent for plan implementation?

Use compose:subagent when subagents are available, since the skill notes it produces significantly higher quality. Use compose:execute when working in a single session, dispatching native child tasks only for independently executable work.

Can plan execution start directly on the main or master branch?

No. The skill explicitly forbids starting implementation on main or master without explicit user consent. It integrates with compose:worktree to create or verify an isolated workspace before executing tasks.

What happens after all plan tasks are completed?

After all tasks complete and pass verification, the skill invokes compose:report to write a human-readable summary of what was built. The report skill then transitions to compose:merge to finish the development workflow.