- Home
- Skills
- Agents & Orchestration
- project plan debate
Works with the AI Agents you already use
project plan debate
Turn project ideas into executable coding or work plans through structured multi-agent debate and critique.
$4.99
project plan debate
Example session with this skill installed
I want to build a decentralized task manager for small teams. Let's start the planning process.
- Read your context and instructions
- Compiled the project plan debate
I have initialized the project debate for your decentralized task manager. I will now guide you through Phase 1 to define the core idea.
Initialized Debate State
| File | Status | Description |
|---|---|---|
plan/_debate/state.md | Phase 1: General | Tracking project progress and gates. |
plan/decisions.md | Active | 0 decisions recorded. |
plan/open-questions.md | Active | Awaiting initial brief responses. |
plan/assumptions.md | Active | Awaiting user-defined unverified beliefs. |
Phase 1: The Idea Brief
To start, I need to understand the problem this solves. Please answer the following:
The Problem: Describe a specific time a small team struggled with task management where existing tools failed. What was the tangible cost of that failure?
2.
Success Metrics: If this project is "finished" in six months, what is one baseline metric that proves it worked, and what is one condition that defines it as a failure?
3.
Project Type: Is this primarily a
Software project (for Claude Code execution), a
Non-software work plan, or a
Mixed project involving both?
Suggestion: Consider if "decentralized" refers to the data architecture (P2P/CRDTs) or the organizational structure, as this significantly changes the Architecture section in Phase 3.
Next steps
- Answer the three questions above to generate the Idea Brief.
- We will then pass Gate 1 and move to the Critic agent for Phase 2.
- Confirm if I should save this plan folder in your current working directory.
Connects securely to your tools. The creator never sees your data.
About this skill
Vague project specs lead to technical debt, scope creep, and wasted development cycles. This skill enforces a rigorous, multi-stage debate process to pressure-test ideas before a single line of code is written. It replaces guesswork with a structured requirements gathering framework that results in a production-ready plan folder.
What it does
- Adversarial Analysis uses a critic agent to challenge project feasibility, licensing, and existing market alternatives.
- Structured Logic Gates prevents moving to the next phase until specific criteria (Idea Brief, Go/No-Go decision, Reviewer approval) are met.
- Dual Plan Generation produces either a
CLAUDE.mdcoding plan for autonomous agents or a human-readable work plan for teams. - Zero-Guessing Requirement logs all unverified beliefs into an assumptions tracker with specific owners and validation dates.
- Multi-Agent Orchestration utilizes specialized sub-agents for scope, architecture, security, and deployment sections.
How it works
- General Debate involves a high-level questioning phase to define the problem, success metrics, and project type.
- Critique & Premortem forces you to justify the project against a critic agent and imagine failure scenarios to identify hidden risks.
- Detailed Sectioning runs specialized agents through 6-8 domain-specific modules (e.g., Security, Architecture) to extract technical specs.
- Draft & Review executes a loop where a reviewer agent finds inconsistencies until the plan is verified as READY.
Frameworks & tools
The skill is optimized for Claude Code and Cursor. It generates specific documentation for CLAUDE.md workflows and utilizes sub-agent architectures for parallel processing.
Why this beats prompting it yourself
Standard prompts often hallucinate missing details or agree with flawed logic to be helpful. This skill uses a formal "critic" role and mandatory state-tracking files to ensure every project decision is explicitly confirmed by you, not assumed by the LLM.
Use cases
- Defining the full tech stack and roadmap for a new SaaS MVP.
- Auditing a project idea for licensing and feasibility before allocating budget.
- Generating a task-breakdown that a coding agent can execute without further instruction.
Known limitations
Requires a plan folder to be initialized on disk for state persistence. Agents cannot talk to the user directly, only via the orchestrator.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Agents & Orchestration.
Requires local disk access for plan folder state. Agents must be run in sequence for Phase 1 and 3. External research is limited to the critic agent phase.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 8 days ago
- Software
Git