- Home
- Skills
- Agents & Orchestration
- Clarify Requirements
Clarify Requirements
Prevents rework by mapping vague requests into structured decision trees before implementation starts.
$5
Works with the AI Agents you already use
Clarify Requirements
Example session with this skill installed
I want to build a CLI tool for managing my team's internal documentation. Help me figure out how to start.
- Read your context and instructions
- Compiled the clarify requirements
I have analyzed your request for an internal documentation CLI. I've automatically checked your current directory and found no existing config files. Before I draft the architecture, I need to resolve the following foundational decisions.
Project Requirements Clarification
❓ Q1 - Storage Strategy: Where should the documentation source of truth live?
- A) Local Markdown files in the project repo (Git-based).
- B) A centralized database (PostgreSQL/Supabase).
- C) An external CMS via API (Notion/Sanity).
➡️ Recommendation: Option A. It's easiest for developers to maintain via PRs and stays version-controlled.
❓ Q2 - Distribution Method: How will the team install and run the tool?
- A) Global NPM/Python package.
- B) A standalone binary (Go/Rust).
- C) A Dockerized environment.
➡️
Recommendation: Option B. Single binaries are faster for internal tools and avoid dependency hell on teammate machines.
❓ Q3 - Search Capabilities: How sophisticated does the lookup need to be?
- A) Simple grep-style keyword matching.
- B) Full-text search engine (FlexSearch/Algolia).
- C) AI-powered semantic search (Embeddings).
➡️ Recommendation: Option A for an MVP, graduating to Option B if the doc volume exceeds 500 files.
Next steps
- Confirm if you agree with the recommended answers (A, B, A).
- Specify if there are any mandatory security protocols for internal data access.
- I will then map the sub-decisions for the chosen storage and distribution path.
Connects securely to your tools. The creator never sees your data.
About this skill
Vague project briefs lead to wasted development cycles and endless rework. This skill stops you from guessing by systematically mapping a request into a decision tree, surfacing only the questions that require human intent before a single line of code is written.
What it does
- Decision Tree Mapping breaks down complex tasks into logical dependencies and sub-decisions.
- Frontier Questioning identifies and asks only the next set of questions that have all prerequisites met.
- Automated Fact-Checking ensures the agent searches files, paths, and environments first instead of bothering you with basic context.
- Tiered Recommendations provides a specific suggested answer for every question to speed up the approval process.
- Consensus Locking prevents development from starting until every critical decision path is confirmed.
How it works
- Analyze the request to build a dependency map of necessary decisions.
- Scan the local environment to resolve any factual variables automatically.
- Prompt the user with a batch of "Frontier" questions—those at the current edge of the decision tree.
- Iterate based on user feedback to unlock the next layer of dependent questions.
Frameworks & tools
This skill is framework-agnostic and functions within any CLI or IDE environment where an agent has read access to the local filesystem.
Why this beats prompting it yourself
Generic agents often dive into implementation with hidden assumptions, leading to "hallucinated" project structures. This skill enforces a structured inquiry phase that ensures the agent only asks what it cannot discover itself, reducing total interaction cycles.
Use cases
- Defining the architecture for a new multi-service system.
- Clarifying performance vs. readability trade-offs in a refactor.
- Gathering specific styling and behavior preferences for a UI component.
Known limitations
Does not support executing code during the questioning phase. Requires a clear initial user intent to begin mapping the decision tree.
Requires an initial prompt with at least one vague requirement. Cannot proceed if the agent is denied read-access to the local environment for fact-checking.
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 any payment.
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 1 month ago
Permissions
No special permissions declared or detected