agentleFS
maxatworkmaxatwork/skills/skills/penpot-canvas/SKILL.md

penpot-canvas

Skill0 starsChanged 8 days agoMITView on GitHub

Use when creating, opening, editing, inspecting, or previewing designs in a user-configured Penpot instance, or when a design should serve as a visual spec. Works through Codex's built-in browser and the Penpot MCP server.

The file

Files
SKILL.md

Penpot canvas

Resolve the Penpot URL from a URL supplied for the current task, then PENPOT_URL, then ${XDG_CONFIG_HOME:-$HOME/.config}/skillbox/penpot.url. The config file contains one URL. For the environment variable or file, use an HTTP(S) origin (scheme, host, optional port) without credentials, path, query, or fragment; trim surrounding whitespace and a trailing slash. If none is set, ask the user for the Penpot URL. Use that instance for browser navigation and design links. Configure the Penpot MCP client for the same instance separately; this URL setting controls browser navigation only.

Work in Codex's built-in browser (iab), in the background by default. Do not open the user's main browser as a dependency. Show the browser only when the user asks to view or interact with the design; return a direct design URL and a preview when useful.

Open or create a design

Use the Penpot dashboard in the built-in browser to find a file by name, open a specified file, or create a new one in the appropriate project. Confirm the file name and page before editing. Do not assume a previously focused file is still the intended target.

The browser profile may retain a Penpot login. If it has expired, ask the user to sign in themselves in the built-in browser; do not type, autofill, or submit credentials on their behalf. Never reveal, extract, copy into a script, or log the password or browser cookie.

Connect and edit

Use the configured Penpot MCP tools when available. In an open design file, connect Penpot's MCP plugin through File → MCP Server → Connect or the equivalent visible control. The MCP server acts on the focused page of the active Penpot tab; a dashboard tab alone is insufficient. If the connection is absent after a browser restart or tab sleep, reopen the design and reconnect on demand. Keep the built-in browser tab for a later turn with markHandoff() when continued work is expected.

Before a write, use a read-only MCP call to confirm the connected file and page. Recheck after browser navigation or a page switch. Start with high_level_overview when the MCP API has not yet been read in the task, and use penpot_api_info for the relevant API surface. Then use execute_code for the requested edit. Inspect the changed object or page afterward. When useful, use export_shape or a browser screenshot to give the user a visual preview.

If the MCP menu is missing or no file is connected, verify the visible workspace and configured MCP availability. Report the specific failure; do not silently edit a different active file. The official MCP plugin depends on a live open file and cannot stay connected after its browser tab is closed or unloaded.

Authentication boundaries

The configured MCP userToken authenticates the MCP connection. It is not a Penpot personal access token for protected backend RPC calls. Use the signed-in browser UI for file discovery and creation. Only use backend RPC when a separately issued Penpot personal access token is available and the deployment has enable-access-tokens configured; never reuse the MCP key as an RPC token. Do not put either token in skill files, design URLs, screenshots, or responses.

Penpot references: MCP workflow, API tokens and RPC, self-hosted flags.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.