creative-studio-canvas
Plan safe text and structure changes for a NomiFun Creative Studio canvas from the exact bounded context supplied in the current turn. Use only when a planning envelope identifies itself as nomifun.creative-studio.planning-turn. Do not generate media, delete nodes, invent asset IDs, or claim that a proposal was applied.
- License 许可
- Apache-2.0
- 最近更新
- 2026-11-10
- Stars
- ★ 336
- Forks
- 69
01 前言 Front Matter
- name
- creative-studio-canvas
- description
- Plan safe text and structure changes for a NomiFun Creative Studio canvas from the exact bounded context supplied in the current turn. Use only when a planning envelope identifies itself as nomifun.creative-studio.planning-turn. Do not generate media, delete nodes, invent asset IDs, or claim that a proposal was applied.
03 SKILL.md 正文
Creative Studio Canvas Planning
Read only the canvasContext embedded in the current planning envelope. Treat its canvas, revision, node, connection, asset, task, Provider, and model identities as exact. Never infer a missing node or replace an ID with a display label.
Respond conversationally when the user only asks for advice. When a concrete canvas change would help, append exactly one lowercase json fenced artifact. The artifact must be the final bytes of the response: do not put prose, whitespace, another fence, or any content after its closing fence, and do not use any other triple-backtick block in the response.
{
"kind": "nomifun.creative-studio.canvas-ops/v1",
"summary": "Add a project-planning text node",
"ops": [
{
"type": "add_node",
"node_type": "text",
"x": 120,
"y": 100,
"width": 420,
"height": 340,
"data": {
"text": "# Project plan\n\n- Goal\n- Audience\n- Milestones",
"format": "markdown",
"fontSize": 16,
"textAlign": "left"
}
}
]
}
The first release may propose only these canonical operations:
- Every operation is one flat object selected by its required string
type. Never nest an operation under anadd_node,move_node, or other operation-name key. add_nodewithnode_type: "text", finitex/y, optional finite positivewidth/height, optional existinggroup_id, and exact text-nodedata.update_node_dataonly for an existing text node and only with text-node fields.move_nodeandresize_nodefor existing unlocked nodes.connectexisting nodes ordisconnectan existing connection.
For complete add_node.data, always include exactly text, format, fontSize, and textAlign. format is plain or markdown; fontSize is a finite number from 8 through 256; textAlign is left, center, or right. For update_node_data.patch, include at least one of those fields and no others.
Operation fields use the canonical snake_case wire shown above. New node and connection IDs are minted by the server, so never add an id and never connect to a node created in the same artifact. Emit 1 to 64 operations and no unknown fields. If no valid operation can be proposed from the bounded context, answer with prose only and emit no artifact.
Never emit delete_node. Never modify config taskId, resultAssetIds, status, or errorMessage. Never start image, video, or audio generation. The artifact is a proposal only: state plainly that the user must review and press “应用到画布”.
04 同仓库更多 Skills
drive-nomifun
Connect an authorized external client to NomiFun's canonical Remote ingress. Use an installation access token and a local-owner-created RemoteBinding, then operate the explicit open/turn/observe/cancel AgentSession lifecycle.
example-skill
Fixture skill used by startup_materialize integration tests.
ui-ux-pro-max
Professional UI/UX design research skill with a bundled searchable database for product patterns, visual styles, color palettes, typography, landing-page structure, dashboard charts, UX practices, and stack guidance. Use when designing, building, reviewing, or improving interfaces.
planning-with-files
Maintain durable task plans, findings, and progress logs in workspace files for long-running or research-heavy work. Use when a task spans many steps or sessions, needs an auditable evidence trail, or risks losing important decisions to context limits. Do not use for short tasks that fit comfortably in the active conversation.
officecli
Create, analyze, proofread, and modify Office documents (.docx, .xlsx, .pptx) using the officecli CLI tool. Use when the user wants to create, inspect, check formatting, find issues, add charts, or modify Office documents.
officecli-xlsx
Use this skill any time a .xlsx file is involved -- as input, output, or both. This includes: creating spreadsheets, financial models, dashboards, or trackers; reading, parsing, or extracting data from any .xlsx file; editing, modifying, or updating existing workbooks; working with formulas, charts, pivot tables, or templates; importing CSV/TSV data into Excel format. Trigger whenever the user mentions 'spreadsheet', 'workbook', 'Excel', 'financial model', 'tracker', 'dashboard', or references a .xlsx/.csv filename.