pavilio-writing-plans

Writes contract-style implementation plans with tasks, acceptance criteria, and test names from approved specs.

Updated Apr 19, 2026
One-click install
npx skills add https://github.com/gmotyl/pavilio --skill pavilio-writing-plans-gmotyl
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: pavilio-writing-plans
Source: https://github.com/gmotyl/pavilio/tree/main/skills/pavilio-writing-plans
Command: npx skills add https://github.com/gmotyl/pavilio --skill pavilio-writing-plans-gmotyl

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Turning an approved design into an executable implementation plan often produces vague, bloated documents with pre-written code that goes stale before execution. This Skill generates a bite-sized, contract-style plan that defines what must be true when each task is done, delegating the how to the implementer. ## Core Features & Use Cases - Contract-Style Tasks: Each task specifies exact file paths, symbol anchors, WHEN/THEN acceptance criteria, and test names instead of pre-written code. - OpenSpec Integration: Saves the plan as tasks.md inside the change's openspec/changes/<change-id>/ directory alongside proposal.md and design.md. - Test-First Granularity: Enforces a bite-sized workflow per task — write failing tests, minimal implementation, verify pass, then commit. - Use Case: After a design spec is approved, invoke the Skill to produce a tasks.md contract that executor sub-agents can implement task-by-task with review checkpoints. ## Quick Start Ask the agent to run /pavilio-writing-plans to turn the approved design spec into an implementation contract saved as tasks.md in the change directory.

Frequently Asked Questions about pavilio-writing-plans

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

FAQPage Schema
How do I turn an approved design spec into an implementation plan?▼

Invoke /pavilio-writing-plans after the design is approved. It resolves the change's OpenSpec backend, then writes a tasks.md contract with files, WHEN/THEN acceptance criteria, and test names for each bite-sized task.

What should an implementation plan for AI coding agents contain?▼

A plan should contain exact file paths, symbol-anchored modify references, WHEN/THEN acceptance criteria, test names, and per-step commands with expected outcomes. It should avoid pre-written implementation code, which goes stale before execution.

Where are implementation plans saved in an OpenSpec workflow?▼

Plans are saved as tasks.md inside the change's openspec/changes/<change-id>/ directory, alongside proposal.md and design.md. Multi-repository changes write a tasks.md under each repository's own change dir.

When should code be included in an implementation plan?▼

Include code only when the code itself is the decision being locked in: public API signatures, exported types, wire formats, schemas, exact config values, or genuinely tricky algorithms. Mark these as Contract blocks inside the task.

What happens if a plan is executed in a later session?▼

A plan is fresh only in the session that wrote it. Dispatched later, it is presumed stale and must pass the staleness check in the pavilio-execute-plan step 1 before implementation begins.