plan-executer

Executes approved step-by-step development plans from docs/current-plan.md on dedicated git branches.

5|Updated Jun 22, 2026
One-click install
npx skills add https://github.com/A158Coke/WotbTools --skill plan-executer-a158coke
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: plan-executer
Source: https://github.com/A158Coke/WotbTools/tree/main/.agents/skills/plan-executer
Command: npx skills add https://github.com/A158Coke/WotbTools --skill plan-executer-a158coke

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Turning an approved development plan into actual code changes is error-prone: steps get skipped, verification is forgotten, scope creeps beyond the plan, and changes land directly on main. This Skill operationalizes the approved plan in docs/current-plan.md by executing each step in order with mandatory verification, branch isolation, and a closing review pass. ## Core Features & Use Cases - Branch-per-task workflow: Derives a kebab-case branch name from the plan topic, syncs main, and confines all changes to that branch. - Stepwise execution with verification: Reads the plan's status table, implements each pending step, runs the corresponding tests/build, and updates plan status after every step. - Blocking and scope control: Stops and reports when encountering unconfirmed blocking items, plan/code conflicts, or out-of-scope changes instead of silently expanding scope. - Acceptance and review closure: Validates every acceptance criterion against concrete tests, runs affected regression tests, updates the changelog, invokes the review-with-docs audit, and opens a PR via gh when zero blockers remain. - Use Case: After plan-designer produces an approved development plan, tell the agent to execute docs/current-plan.md and receive a reviewed, tested pull request on a dedicated feature branch. ## Quick Start Ask the agent to execute docs/current-plan.md according to the approved plan and report progress step by step.

Frequently Asked Questions about plan-executer

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

FAQPage Schema
How do I execute an approved development plan step by step with an AI agent?

Place the approved plan in docs/current-plan.md and instruct the agent to execute it. The Skill reads the full plan, creates a task branch, implements each pending step with verification, updates the plan status table, and finishes with a review pass and pull request.

What is the difference between plan-designer and plan-executer?

plan-designer produces the development proposal document before coding, while plan-executer implements an already approved plan from docs/current-plan.md. They are complementary: design first, approve, then execute.

Does plan-executer commit changes directly to the main branch?

No. It always syncs main, creates a kebab-case branch named after the plan topic with a feat/fix/refactor/chore prefix, and confines all changes to that branch. Pull requests are opened via gh only after the review pass reports zero blockers.

What happens when a plan step is blocked or conflicts with the code?

Execution stops and the Skill reports the blocking item with file and line references, offering options to the user. It never silently expands scope, overrides the approved plan, or proceeds past unconfirmed blocking items.

Does plan-executer run the full test suite after finishing?

No. It runs only the affected regression tests for the task; repository-level full validation is delegated to PR CI. Full test runs happen only in documented exception cases with a declared scope and rationale.