plan

Writes self-contained implementation plans with verifiable completion criteria and explicit constraints.

Updated May 3, 2026
One-click install
npx skills add https://github.com/spikelab/multiplai-cc-mktplace --skill plan-spikelab
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: plan
Source: https://github.com/spikelab/multiplai-cc-mktplace/tree/main/plugins/multiplai-dev/skills/plan
Command: npx skills add https://github.com/spikelab/multiplai-cc-mktplace --skill plan-spikelab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Plans that only describe steps leave the executing agent to decide when work is done, causing scope drift and premature declarations of completion. This Skill produces plan files that define done through verifiable criteria, explicit out-of-scope rules, and stop-and-ask gates that survive context compaction and session restarts. ## Core Features & Use Cases - Completion contract in every plan: Each plan includes numbered Done means criteria checkable by a specific command or observation, plus Constraints and stop-and-ask gates. - Recon-grounded writing: Verifies every referenced path, filename, branch, and command against the actual repo before writing, and flags non-homogeneous scope for the user to split. - Fresh-session executability: Self-tests the plan so a session with zero conversation history can execute it from the file alone, making it directly consumable by goal runners or buildme. - Use Case: Ask for an implementation plan for a multi-file refactor; receive a dated plan file in the project's plans directory with ordered work items, exclusions, and a verification section with exact commands. ## Quick Start Ask the agent to write an implementation plan for your next feature, for example by saying write a plan for adding OAuth login to the app.

Frequently Asked Questions about plan

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

FAQPage Schema
How do I write an implementation plan an AI agent can execute?

Define numbered Done means criteria that are each verifiable by a specific command or observation, list explicit constraints and out-of-scope items, and make the plan self-contained so a fresh session needs no conversation history. Verify every referenced path and command exists before writing.

What should a good implementation plan template include?

Include Objective, Context, Outcome, ordered Work items with named deliverables, Constraints and out-of-scope rules with stop-and-ask gates, verifiable Done means criteria, and a Verification section with exact commands and expected outputs. No section may be omitted.

When should I use a separate goal document instead of a plan?

Use a separate goal document only when the goal aggregates several plans, such as multi-repo initiatives or parallel plan tracks sharing a frozen contract. For a single plan, a second document duplicates content and the two artifacts drift out of sync.

Why do agents drift out of scope when executing plans?

Scope drift happens when a plan describes how but never defines done or exclusions, so the agent judges completeness itself. Writing explicit constraints, exclusions, and stop-and-ask gates into the plan file keeps boundaries intact across context compaction and restarts.

Can a plan file be used by an autonomous goal runner?

Yes. The Outcome, Done means, and Constraints sections form the goal contract, and the Work items provide the decomposition. The same file also seeds the proposal stage of buildme without any additional wrapper document.