More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    Spec Driven Planning

    1

    Generate human-verified spec, plan, and task files to guide coding agents before development begins.

    Free

    Security scanned

    spec-driven-planning

    Example session with this skill installed

    /spec-driven-planning. I want to build a simple CLI tool in Go that fetches stock prices from an API and saves them to a CSV file. It sounds like a Simple tier.

    • Read your context and instructions
    • Compiled the spec-driven-planning

    I have processed your request for the Go CLI tool.

    But I would recommend you to answer a few questions to enhance the spec.md file.

    Q1. How would you like to ...?

    Your Answer...

    Q2. I have kept these dependencies, do you want to change or remove anything?

    Your Answer...

    Based on our interaction, I have generated the project blueprint.

    spec.md

    
    # Project Spec: Stock-to-CSV CLI
    
    ## Goal
    A Go-based command line interface that retrieves current stock prices for given tickers and persists the data to a local CSV file.
    
    ## Requirements
    - Support fetching data for one or multiple tickers via CLI flags.
    - Integrate with a public stock market API (e.g., Alpha Vantage or Yahoo Finance).
    - Output a CSV file named `prices.csv` with columns: Ticker, Price, Timestamp.
    - Handle API rate limiting by implementing a simple retry or exit strategy.
    
    ## Non-Functional Requirements
    - Language: Go 1.21+
    - No external database; local file storage only.
    - Minimal external dependencies.
    
    ## Acceptance Criteria
    - Running `stock-cli --tickers AAPL,MSFT` creates or appends to `prices.csv`.
    - The CSV contains valid data points for all requested tickers.
    - Errors (like invalid tickers) are logged to stderr without crashing the tool.
    

    Next steps

    1. Review the generated spec.md in your project root.
    2. If satisfied, you can now hand this file to your coding agent to begin the implementation.
    3. If you need to add auth or a database later, we can run this again to upgrade the project to Medium tier.
    spec.md

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Handing a vague prompt to a coding agent leads to scope creep and inconsistent code. This skill enforces a "blueprint first" workflow, generating verified documentation before any code is written.

    What it does

    • Tier-based scoping determines if a project is Simple, Medium, or Complex to match the documentation depth.
    • Guided elicitation uses structured question sets to extract requirements without missing critical details.
    • Spec generation creates spec.md to define the "what" and "why" of the project.
    • Architecture planning produces plan.md for Medium and Complex projects to map out modules and logic.
    • Task breakdown generates tasks.md with acceptance criteria and traceability for Complex builds.

    How it works

    1. Invoke the skill using the /spec-driven-planning command or a direct request.
    2. Select a project tier based on complexity examples provided by the agent (or ask the agent to propose the tier itself)
    3. Answer targeted questions conversationally to define features, constraints, and architecture.
    4. Review and approve each generated Markdown file at specific checkpoints before moving to the next.

    Frameworks & tools

    This tool outputs standard GitHub-flavored Markdown files. It works with any IDE, version control system, or coding agent that can read .md files in a local directory.

    Why this beats prompting it yourself

    Manual prompts often lack clear "done" criteria and architectural boundaries. This skill forces human-in-the-loop review of the plan, preventing the agent from making silent, incorrect assumptions about your stack or logic. This also prevents unnecessary iteration loops after the agent shows its tasks to you.

    Use cases

    • Drafting a technical spec for a new feature before handoff to a junior dev or agent.
    • Scoping a complex multi-stage application that requires a database and auth.
    • Standardizing project documentation across a team using consistent Markdown templates.

    Known limitations

    Does not write the actual application code. Output is strictly limited to Markdown documentation. Requires manual invocation by the user.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Prompt & Skill Engineering.

    Does not write application code. Limited to Markdown documentation. Requires explicit manual invocation to trigger.

    How to use

    1. 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. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 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

    Free to use

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 20 days ago

    • Free with an Agensi account

    Nothing to install. It runs in your AI agent.

    Optional
    Node.js

    Permissions

    Read Files
    Write Files
    Listed20 days ago

    What's inside

    Frequently Asked Questions

    This skill eliminates 'hallucinated' project structures by enforcing a human-verified blueprint. By generating specific spec, plan, and task files before any code is written, it prevents agents from making incorrect assumptions about your architecture or project scope. The aim is to define what "Good" is beforehand, to avoid going into iterative loops.

    The skill outputs standard GitHub-flavored Markdown (.md) files. This makes it compatible with any coding agent (like Claude Code, Cursor, Windsurf, or GitHub Copilot) or IDE that can ingest local documentation to guide its development process.

    Tiers range from 'Simple' (generating only a spec.md) to 'Complex' (which includes full architecture plans and a tasks.md with acceptance criteria). This ensures you aren't bogged down by over-documentation for small scripts, while providing deep traceability for full-stack builds.

    No, this skill is strictly a planning and architecture tool. It produces the technical roadmap and task list that you then provide to a coding agent or developer to execute the actual implementation.

    New in Prompt & Skill Engineering

    Free