executing-plans

Executes written implementation plans in task batches with review checkpoints.

1|Updated Feb 10, 2026
One-click install
npx skills add https://github.com/stefanfaur/roach-marketplace --skill executing-plans-stefanfaur
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: executing-plans
Source: https://github.com/stefanfaur/roach-marketplace/tree/main/roach/skills/executing-plans
Command: npx skills add https://github.com/stefanfaur/roach-marketplace --skill executing-plans-stefanfaur

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Executing a written implementation plan manually risks skipped steps, missed verifications, and uncontrolled deviations. This Skill enforces a disciplined batch-execution workflow with checkpoints so plans are implemented faithfully and reviewed incrementally. ## Core Features & Use Cases - Batch Execution with Checkpoints: Executes tasks in batches (default 3), then pauses to report results and await feedback before continuing. - Task Persistence: Loads and syncs task state from a <plan-path>.tasks.json file co-located with the plan, enabling session resumption. - Deviation Rules: Distinguishes auto-fixable local issues (typos, wrong import paths) from cross-task changes that must be escalated to the user. - Use Case: You have a written implementation plan from a prior planning session. Use this Skill to load it, execute the first batch of tasks with verifications, report results, and iterate until completion. ## Quick Start Use the executing-plans skill to implement the plan in thoughts/shared/plans/my-feature.md, executing tasks in batches and reporting after each batch.

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 assistant?▼

Load the plan file, review it critically, then execute tasks in batches of three, marking each in_progress and completed as you go. After each batch, report what was implemented with verification output and wait for feedback before continuing.

How to resume plan execution in a new session?▼

The skill checks for a <plan-path>.tasks.json file co-located with the plan. If found, it recreates tasks from the JSON, restores blockedBy relationships, and resumes from the first pending or in_progress task.

When should plan execution stop and ask for help?▼

Stop immediately when hitting a blocker mid-batch, when the plan has critical gaps, when instructions are unclear, or when verification fails repeatedly. Ask for clarification rather than guessing or forcing through blockers.

What deviations are allowed when executing a plan?▼

Auto-fix local issues like wrong import paths, typos, missing error handling, or broken references, and note them in the batch report. Escalate anything with cross-task implications, such as new schemas, different libraries, or weakening existing tests.

Can this skill run in plan mode?▼

No. The skill explicitly forbids EnterPlanMode and ExitPlanMode because it executes an existing plan in normal mode and needs unrestricted Write and Edit tools for implementation.