spec driven development

    1

    Enforce a gated, four-phase workflow to define specs, maps, and tasks before writing code.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+22 more

    spec driven development

    Example session with this skill installed

    I need to add a subscription billing system to our existing Next.js app. It needs to handle Stripe, tiered pricing, and seat-based billing. Give me a spec.

    • Read your context and instructions
    • Compiled the spec driven development

    I have analyzed the request and identified that this touches multiple capabilities (identity, payments, and usage tracking). I will start by surfacing my assumptions before providing the spec document.

    ASSUMPTIONS I'M MAKING:

    1. We are using Stripe Checkout/Billing rather than a custom PCI-compliant flow.
    2. Users must be authenticated before accessing billing features.
    3. Subscription status will be mirrored in our local PostgreSQL database via webhooks.
    4. The frontend is built using React Server Components.
      → Correct me now or I'll proceed with these.

    Spec: Subscription Billing System

    Objective

    Implement a tiered subscription system using Stripe to allow per-seat billing and automated invoicing for SaaS customers.

    Tech Stack

    • Next.js 14 (App Router)
    • Stripe Node SDK
    • Prisma ORM (PostgreSQL)
    • Tailwind CSS

    Commands

    Build: npm run build
    Test: npm test
    Stripe Webhook Listen: stripe listen --forward-to localhost:3000/api/webhooks

    Project Structure

    src/app/billing → Billing dashboard and checkout pages
    src/app/api/webhooks → Stripe webhook handlers
    src/lib/stripe → Stripe client and helper functions

    Success Criteria

    • User can select a plan and complete Stripe Checkout.
    • Database updates user subscriptionStatus upon successful invoice.paid webhook.
    • Access to /dashboard is revoked if subscriptionStatus is not active.

    Next steps

    1. Review the assumptions above.
    2. Approve this spec to move to the Plan phase.
    3. Provide the existing schema.prisma file if you want the spec to include specific data models.

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

    About this skill

    Avoid the trap of coding based on assumptions. When requirements are vague or a single request touches multiple systems, jumping straight to implementation leads to rework and architectural debt. This skill enforces a structured, four-phase gated workflow to ensure you and your agent are aligned before a single line of code is written.

    What it does

    • Decompose complex requests by creating a capability map that identifies distinct modules and their dependencies.
    • Surface assumptions by listing technical and environmental premises for human validation before proceeding.
    • Define concrete specs covering objectives, stack, directory structure, and executable commands.
    • Establish guardrails with a three-tier boundary system (Always, Ask First, Never) to prevent unauthorized changes.
    • Reframe requirements by translating vague user requests into specific, testable success criteria.

    How it works

    1. Scope Check. The agent detects if the request contains multiple testable capabilities and proposes a module map.
    2. Specify. The agent generates a SPEC.md document covering six core areas and waits for human approval.
    3. Plan. A technical implementation strategy is created in tasks/plan.md with dependency-graph mapping.
    4. Tasking. The plan is broken into discrete, session-sized tasks with clear verification steps in tasks/todo.md.

    Frameworks & tools

    This skill is framework-agnostic and works with any stack including React, Node.js, Python, and PostgreSQL. It integrates with existing incremental-implementation and test-driven-development workflows.

    Why this beats prompting it yourself

    Generic prompts often lead to "hallucinated" requirements or silent assumptions that break your architecture. This skill forces a gated approval process, ensuring the agent never builds the wrong thing for 15 hours when a 15-minute spec could have caught the error.

    Use cases

    • Starting a greenfield project where the folder structure and tech stack are undecided.
    • Adding a major feature that spans multiple modules, like adding billing to an identity system.
    • Refactoring legacy code where success criteria must be strictly defined to avoid regressions.

    Known limitations

    Not intended for single-line fixes or trivial typo corrections. Requires manual human approval at each phase gate to advance to implementation.

    How to use

    Skills on Agensi run inside your AI agent. Copy the prompt, paste it into your agent, and it connects to Agensi and installs this skill for you. If a payment is needed, your agent shows it to you first.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill passes our trust checks below.

    Security scanned

    Passed our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scanned

    Verified clean 2 days ago

    • Free to download with an account

    Listed2 days ago

    What's inside

    Frequently Asked Questions