More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Brainstorming: Ideas Into Designs

    1

    Turn a rough idea into a design worth building — lock intent in writing, generate real options, stress-test the favorite before any code.

    Free

    1 useSecurity scanned
    Brainstorming: Ideas Into Designs

    Brainstorming: Ideas Into Designs

    Example session with this skill installed

    I have a single-player fishing game that sells steadily, and players keep asking for co-op. I like the idea but I have no clue what it would take, and I do not want to rebuild the whole game around multiplayer. Help me figure out what we should actually build before I write any code.

    • Read your context and instructions
    • Compiled the brainstorming

    Intent note (correct me where this is wrong).
    Outcome: two players can fish together in one session without the single-player game getting worse. Who it is for: existing players who want to play with one friend, not strangers. Success looks like: a pair finishes a full in-game day together, catches saved correctly, no sync bug bad enough to make either quit. Constraints: solo developer, no server budget, saves are local files, game stays playable offline.
    You said: players ask for co-op and you do not want a rebuild. I am assuming: co-op means 2 players, one hosts, same platform, progress lives in the host's save. If any of that is wrong, the design changes.

    Depth chosen: Brief. This touches saving, session flow, and the core loop, too much for a napkin answer, but it is not a new system with its own data model, so a full spec is overkill today.

    The options.

    1. The boring one: host-and-guest sessions over the networking your engine already ships. The host's game is the truth; the guest joins the host's lake and save. Cost: weeks, mostly session flow and save rules. Risk: moderate and known. Rules out: easy matchmaking later.
    2. The ambitious one: dedicated servers, accounts, persistent shared lakes. Cost: months plus running costs, for a solo dev with no server budget. Rejected because it breaks a stated constraint.
    3. The deleted one: no real-time co-op. Add shared goals instead, like a weekly community catch target and giftable bait. Cost: days. Only a partial answer to "play with friends", but the honest fallback if option 1 fails its test.

    Stress test of the favourite (option 1). What breaks first: save ownership. If the guest catches a rare fish, whose collection grows? Decide that before any networking code. The answer here: host's world, host's collection, guest keeps a personal catch log. Smallest proof it could be wrong: a two-laptop spike on one lake, hard-coded connection, one afternoon. If catching does not stay in sync there, option 1 dies cheaply and option 3 becomes the plan. Price paid later: public sessions get harder, which we accept.

    Approval gate. This brief is the deliverable. Nothing gets built until you approve the intent note and option 1 as written. Approving the idea is not approving the brief, and approving the brief is not approval to start coding.

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

    About this skill

    Free starter from my paid packs. Before creative or architectural work, this skill runs a structured conversation: what are we actually building and why, what are the real options (not one option and two strawmen), which one survives stress-testing, and explicit approval of the design before implementation starts. Cheap insurance against building the wrong thing well.

    It produces a decision, not a product: the output is an intent note, options, and a stress test at the depth the work deserves, and it stops at an approval gate instead of implementing anything. It also cannot supply the answers only you have; if the intent is unclear, it will keep asking and labelling assumptions rather than filling gaps silently.

    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 25 days ago

    • Free with an Agensi account

    Permissions

    No special permissions declared or detected

    Listed25 days ago

    What's inside

    Frequently Asked Questions

    A written design brief: the intent with assumptions labelled, at least two real options plus the option of not building it, a stress test of the favourite, and a clear approval point. It does not write code or start the build.

    Yes, a rough idea is the normal input. Expect questions back: who it is for, what success looks like, and what is off the table. The sharper those answers, the sharper the options you get to choose from.

    Any agent that supports the SKILL.md standard, including Claude Code, Cursor, VS Code Copilot, and Gemini CLI. Your agent loads the latest approved version each time you use it, so the process you get is always the current one.

    New in Productivity