executing-plans

Executes written implementation plans in task batches with review checkpoints between batches.

Updated May 22, 2026
One-click install
npx skills add https://github.com/kitfunso/claude-config --skill executing-plans-kitfunso
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: executing-plans
Source: https://github.com/kitfunso/claude-config/tree/main/skills/executing-plans
Command: npx skills add https://github.com/kitfunso/claude-config --skill executing-plans-kitfunso

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Written implementation plans often get executed sloppily: steps get skipped, verifications are ignored, missing dependencies are papered over with guesses, and results are reported as done when they were never verified. This Skill enforces a disciplined execution loop so a plan is implemented exactly as written, with honest reporting at every checkpoint. ## Core Features & Use Cases - Critical Plan Review: Before touching code, it reads the plan and verifies with Glob/Grep that every file, module, and function the plan references actually exists, raising concerns before any edit. - Batched Execution with Checkpoints: Tasks run in batches (default 3), each marked in_progress then completed, followed by a report ending with "Ready for feedback" so a human or architect can review between batches. - Honest Verification Reporting: When no shell is available for tests or git, it says so explicitly instead of claiming passes or commits it never observed. - Use Case: You hand the agent a plan at docs/plans/add-clip-price.md. It loads the plan, implements the task and test exactly as specified, reports what it changed without claiming unrun tests passed, and waits for your feedback before continuing. ## Quick Start Ask the agent to execute the implementation plan at a given file path, for example: "Here is the plan at docs/plans/2026-09-18-add-clip-price.md, please execute it task by task."

Frequently Asked Questions about executing-plans

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 agent?▼

Point the agent at the plan file path and ask it to execute the plan. The executing-plans skill loads the plan, reviews it critically, runs tasks in batches of three, and reports back with "Ready for feedback" between batches for your review.

How to make an AI agent follow a plan exactly without skipping steps?▼

Use a batch-execution workflow that marks each task in_progress, follows every step literally, runs the specified verifications, and marks completion only after checks. This skill enforces that discipline and stops to ask rather than guessing when instructions are unclear.

What happens if the plan references files or modules that do not exist?▼

During plan review, the agent searches the repo with Glob/Grep for every file, module, and function the plan imports or modifies. If any dependency is missing, it raises the concern with you before making any edit instead of inventing a workaround.

Can the agent run tests and commits without a shell tool?▼

No. When no shell is available, the agent explicitly reports that it could not run pytest or git rather than claiming a pass or commit it never observed. It still reports the batch and waits for feedback.

When should the agent stop executing a plan and ask for help?▼

It stops immediately on blockers such as missing dependencies, repeatedly failing verifications, unclear instructions, or critical plan gaps. It asks a concrete clarifying question instead of forcing through or guessing an implementation.