writing-plans

Decompose user goals into verifiable task lists saved to .ohc/plans/.

2|Updated Apr 26, 2026
One-click install
npx skills add https://github.com/iadr-dev/colab --skill writing-plans-iadr-dev
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: writing-plans
Source: https://github.com/iadr-dev/colab/tree/main/skills/writing-plans
Command: npx skills add https://github.com/iadr-dev/colab --skill writing-plans-iadr-dev

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) and assets (resource) components.

What problem does it solve?

Writing plans that are vague, underspecified, or too big leads to wasted cycles, scope creep, and agent churn; this skill converts a goal into a precise, decomposed task list with clear success criteria and an explicit approval gate.

Core Features & Use Cases

  • Deep interview planning: asks 3–5 focused questions to clarify scope, constraints, risks, and existing patterns.
  • Approach selection with tradeoffs: proposes 2–3 approaches and requires user confirmation before committing to the full plan.
  • Task decomposition rules: enforces small, independently verifiable tasks (≤2 hours each) with file scope and agent assignment.
  • Plan persistence + gating: saves plans to .ohc/plans/ using the canonical template and blocks BUILD until the user responds “go”.
  • External spec and traceability support: runs document-intake when needed and requires Sources/Traceability for durable alignment.

Quick Start

Run writing-plans by asking your agent to respond with the keyword "plan" for the work you want done, and confirm the proposed approach when prompted.

Frequently Asked Questions about writing-plans

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

FAQPage Schema
How do I break down a large project into an executable task list?

To break down a large project into an executable task list, this skill conducts a deep interview to clarify scope, then decomposes the work into independently verifiable tasks limited to two hours each with specific file scopes and agent assignments.

What is the best way to prevent scope creep when planning complex software development?

The best way to prevent scope creep during software development planning is to require explicit user confirmation on a selected approach from two to three proposed tradeoff options before committing to the full task decomposition and gating subsequent build actions.

How do I ensure traceability when building from external specification documents?

To ensure traceability when building from external specification documents, this skill runs a document-intake process and embeds mandatory Sources and Traceability fields within the canonical plan template saved to the .ohc/plans/ directory.

Do I need user approval before an agent starts building from a generated plan?

Yes, you need explicit user approval before an agent starts building. The generated plan enforces a workflow gate that blocks all BUILD processes until the user reviews the saved plan and responds with the exact keyword "go".

Can I use this planning method for short tasks that take less than 30 minutes?

You should not use this planning method for short tasks taking less than 30 minutes. The deep interview and approach selection process is specifically designed for building efforts that exceed 30 minutes and require strict task decomposition.

How does an agent clarify requirements before generating a development plan?

An agent clarifies requirements before generating a development plan by asking three to five focused questions targeting scope, constraints, risks, and existing patterns, ensuring the final plan precisely matches the user's initial goal.