Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Figma Handoff Gate

    1

    Gate before a Figma section goes to developers: 17 checks. Every drawn action needs a designed outcome, and add needs a matching remove.

    Free

    2 usesSecurity scanned

    About this skill

    Pre-ship gate for a Figma handoff. It runs 17 checks before a section is declared ready for developers and reports PASS, or FAIL with numbered [BLOCKER] and [WARN] items by node id. Both include an action -> outcome map. Read-only until you approve a fix.

    What it checks:

    • Every drawn action has a designed frame, sheet or state. Add needs a remove, help icons open a designed screen or sheet, destructive actions get a confirm state.
    • Frame spacing, frame-name clearance, caption layout and styling.
    • Dark note cards on light screens, 30 to 100 words each.
    • Spec cards of 3 to 5 bullets, no internal jargon, no AI stuffing.
    • A screen count that matches the screens, and one frame per trigger, bypass and error path.
    • Zero em dashes and curly quotes in the section's text, frame names and descriptions.
    • In files a client or stakeholder will open: no process log, no AI tool names, no visible layout grids.
    • Small elements captured at 2x or more, and the prototype starting point confirmed.

    Requires the figma-canon skill installed next to it (it links to ../figma-canon/references/). Run figma-slop-check first; approved fixes go through figma-preflight.

    Part of figma-maxxing, free and MIT on GitHub: https://github.com/thiagoxikota/figma-maxxing

    The skill is instructions for the agent; nothing in it blocks a write on its own. No node type marks an element as interactive, so unwired controls are found by layer and component names and by eye. The check 16 term sweep needs manual triage (an iOS navigation bar's prompt slot and a place name like Mumbai are listed false positives), and a large page may need the figma_execute timeout raised (default 5000 ms, maximum 30000 ms). No HandoffNote component ships; build one from the figma-canon spec if the file has none. In the README's test on the official Figma MCP server, get_screenshot gave no capture above 1x, while check 17a asks for 2x or more.

    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

    2 uses

    Times used to date

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed 1 month ago

    • Free with an Agensi account

    Requires

    Figma

    Permissions

    Read Files

    Allowed Hosts

    api.figma.com

    File Scopes

    ../figma-canon/references/handoff-format.md
    ../figma-canon/references/ai-slop-signatures.md
    ../figma-canon/references/state-coverage.md
    ../figma-canon/references/plugin-api-anomalies.md
    ../figma-canon/references/plugin-api-core.md

    The skill folder holds only SKILL.md: no scripts. The agent reads 5 Markdown references from the sibling figma-canon skill (handoff-format.md on every check). Figma is reached only through the user's own MCP server (figma_execute, figma_capture_screenshot). The skill itself opens no network connection, reads no environment variables and writes no local files. Approved fixes write to the Figma file through that MCP server, one approved fix at a time, after figma-preflight.

    Listed1 month ago

    What's inside

    Frequently Asked Questions

    Not on its own. The gate reports and changes nothing on the canvas until you approve a fix. You approve each fix before it runs, there is no mass mutation, and an approved fix goes through figma-preflight like any other write.

    The figma-canon skill installed next to it, because the checks read its handoff-format.md and ai-slop-signatures.md references. The detector snippets are Plugin API JavaScript for figma_execute (figma-console-mcp with the Desktop Bridge plugin in Figma Desktop). figma-slop-check runs before this gate, and figma-preflight handles approved fixes.

    PASS lists the section, frame and component counts, a timestamp and the action -> outcome map. FAIL lists numbered [BLOCKER] and [WARN] items, each with a node id and a fix, plus the same map. Every orphan action is a blocker.

    New in Branding & Design

    Free