Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Project Brainstorm

    1

    Grilling-style brainstorming that turns a rough software idea into a complete, standards-based design package.

    Free

    1 useSecurity scanned
    Project Brainstorm

    Project Brainstorm

    Example session with this skill installed

    Illustrative demo request, not a recorded user session.

    Use Project Brainstorm to help me design a habit tracker for college students who struggle to maintain study routines. Challenge the idea before choosing features or a tech stack. Ask questions with recommended answers, but wait for my decisions. Keep the first version feasible for a solo developer using free tools. Do not write application code.

    • Read your context and instructions
    • Compiled the project brainstorm

    Illustrative opening response, not a recorded skill run. This demonstrates the first questioning stage only. No design decisions have been approved and no checker has been run.

    Habit Tracker: First Decision Round

    Before choosing features or a tech stack, let's settle the problem, users and design scope.

    1. Who is the first version for?

    Recommendation: Individual college students managing their own study routines, rather than teachers managing a class.

    Why: One user group keeps the design focused and avoids classroom permissions before they are needed.

    2. What problem should it solve?

    Recommendation: Help students notice missed study sessions and restart their routine, rather than simply maximize streak length.

    Why: This gives the product a clear purpose without assuming streaks are the right motivation.

    3. What would make the prototype worth continuing?

    Recommendation: A student can define a study habit, record progress and understand the past week's activity without help.

    Why: This is a proposed workflow we can turn into acceptance tests, not a claim of measured success.

    4. Which design tier should we use?

    Recommendation: Prototype.

    Why: You want a small first version built by one developer. Test the core idea before designing a larger rollout.

    5. How much research should I do?

    Recommendation: Standard research.

    Why: It is the skill's recommended research depth for a Prototype. No research and Deep research are alternatives. Research has not been performed in this illustrative response.

    6. Which solution direction should we explore?

    Recommendation: Compare a simple check-in tracker, a scheduled study planner and an accountability-based approach before selecting one.

    Why: They address different reasons for inconsistency. Choosing screens or storage now would lock in an approach too early.

    7. What other constraints apply?

    Recommendation: Confirm your deadline, target platform and any required technology. Keep the free-tools and no-application-code constraints you supplied.

    Why: These answers can change the feasible design. I will not invent them or treat a recommendation as your approval.

    Your turn

    Reply with answers by number. Accept or change each recommendation. Unanswered decisions stay open. The complete design package comes after the decisions are settled and confirmed.

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

    About this skill

    Software projects often fail because of unstated assumptions and weak architecture. This skill forces you to settle every design decision before you write a single line of code, transforming a rough idea into a professional design package.

    What it does

    • Decision Grilling walks you through a comprehensive design tree to resolve every ambiguity in your project idea.
    • Researched Recommendations provides a vetted "best option" for every question based on current standards and tools.
    • ISO Standard Requirements generates functional and non-functional requirements (NFRs) mapped to ISO/IEC 25010 characteristics.
    • Risk & Threat Modeling identifies vulnerabilities and maps them to specific mitigations or acceptance records.
    • Automated Validation runs a trace-check script to ensure every requirement has a test case and every decision has a source.

    How it works

    1. Use the /project-brainstorm command followed by your initial software idea or problem statement.
    2. Select a research depth (No research, Standard, or Deep) to determine how much the agent investigates tool options and laws.
    3. Respond to "frontier" rounds of questions where the agent presents decisions, trade-offs, and recommended paths.
    4. Confirm the final design summary once the decision tree is empty and the "improve" loop finishes.

    Frameworks & tools

    This skill produces standard Markdown documentation and Mermaid diagrams. It utilizes a custom Python trace-check script for validation and targets open-source stacks and free tiers for all architectural recommendations.

    Why this beats prompting it yourself

    Generic prompts often miss non-functional requirements like security and scalability or fail to push back on poor design choices. This skill uses a formal design tree and a multi-round "grilling" process to ensure no blind spots remain in your architecture.

    Use cases

    • Drafting a complete SRS and architecture for a new MVP to show stakeholders.
    • Stress-testing a legacy system migration plan against modern security and performance standards.
    • Creating a detailed roadmap and API specification for a distributed engineering team.

    Known limitations

    This skill is strictly for software projects; it does not handle hardware design. It prioritizes free and open-source tools unless a paid tool is explicitly requested.

    Software projects only; hardware design is not supported. Produces design documents, not application code or repository scaffolding. Recommends free/open-source tools and free tiers unless you explicitly choose a paid tool. Standard and Deep research need web access; without it, unchecked facts are marked for verification. Automated trace checks require Python 3.8+; manual checking is supported without Python. The checker catches structural and traceability issues, not every design inconsistency. Other-agent compatibility and end-to-end Agensi delivery remain unverified.

    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 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

    1 use

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 5 days ago

    • Free with an Agensi account

    Requires

    Git

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access
    Environment Variables
    Listed5 days ago

    What's inside

    Frequently Asked Questions

    This skill is designed for any AI agent environment that supports Markdown rendering, Mermaid diagrams, and Python execution for the validation scripts. It works best with agents that have high-context windows to manage the complex design trees generated during the 'grilling' process.

    You will receive the core logic for the Decision Grilling process, the ISO-standard requirement templates, and the custom Python trace-check script used for automated validation. It also includes pre-configured research depth settings to guide the agent's investigation of tools and regulations.

    Unlike generic prompts that often agree with the user, this skill is programmed to 'push back' using a rigorous design tree. It forces you to address non-functional requirements like security and scalability that are frequently overlooked in standard AI chats.

    The skill generates a comprehensive suite of documentation including a Software Requirements Specification (SRS), ISO/IEC 25010 mapped requirements, Mermaid architecture diagrams, and a full Risk & Threat Model. All outputs are provided in standard Markdown for easy integration into your existing workflow or version control.

    No complex installation is required; simply import the skill into your compatible AI agent and use the `/project-brainstorm` command to begin. You can immediately start a session by inputting a rough problem statement and selecting your desired research depth.