canva-edit-design
Make edits to an existing Canva design — change or fix text, replace/insert/delete images and videos, reformat text (size, weight, style, color, alignment, lists, line height), reposition or resize elements, and update the title. Use when the user wants to change, edit, update, fix, translate, replace, or reformat content in a specific Canva design. This is the safe edit engine that other Canva skills (e.g. implement-feedback) build on.
- License 许可
- MIT
- 最近更新
- 2026-09-22
- Stars
- ★ 0
- Forks
- 0
01 前言 Front Matter
- name
- canva-edit-design
- description
- Make edits to an existing Canva design — change or fix text, replace/insert/delete images and videos, reformat text (size, weight, style, color, alignment, lists, line height), reposition or resize elements, and update the title. Use when the user wants to change, edit, update, fix, translate, replace, or reformat content in a specific Canva design. This is the safe edit engine that other Canva skills (e.g. implement-feedback) build on.
03 SKILL.md 正文
Canva Design Editing
The canonical, safe way to apply edits to an existing Canva design. Every Canva skill that mutates a design should follow this exact protocol: start a transaction → perform operations → commit (with approval). Changes are draft-only until committed and are PERMANENTLY LOST if not committed.
The Transaction Protocol (always these steps, in order)
Canva:start-editing-transaction— pass thedesign_id. Remember the returnedtransaction_idand thepagesarray; both are required by later calls. ALWAYS show the user the thumbnail(s) returned here.Canva:perform-editing-operations— apply edits. Pass thetransaction_id, thepagesarray from the previous response, thepage_indexof the first page being changed, and anoperationsarray. Batch multiple operations into a single call wherever possible.Canva:commit-editing-transaction— save. See the approval gate below. After committing, thetransaction_idis invalid; a new edit needs a new transaction.Canva:cancel-editing-transaction— discard the draft instead of saving (e.g. the user rejects the preview, or you opened a transaction only to inspect the design).
Capabilities — what the API CAN and CANNOT do
CAN (operations on perform-editing-operations)
- Text content:
replace_text(whole element),find_and_replace_text(substring) - Text formatting (
format_text): font size, weight (normal/bold), style (normal/italic), color, alignment, line height, underline, strikethrough, links, list level/marker - Media:
update_fill(replace image/video),insert_fill(add image/video),delete_element - Layout:
position_element,resize_element - Metadata:
update_title - Autofill mapping:
update_autofill_field(fixed-page designs only)
CANNOT
- Change font family/typeface (only size, weight, style)
- Add new text elements (you can only insert media, not new text boxes)
- Change background colors or gradients
- Add, remove, or reorder pages/slides
- Modify animations, transitions, or element opacity (except on newly inserted fills)
- Group/ungroup elements, or restyle shapes (only text inside shapes is editable)
When a requested change is in the CANNOT list, tell the user it must be done manually in the Canva editor — don't attempt a workaround.
Responsive pages — restricted operation set
Some pages come back marked is_responsive: true. On those pages, ONLY these operations are allowed:
update_title, replace_text, update_fill, delete_element, find_and_replace_text.
Before calling perform-editing-operations, check the pages array. If any operation targets a responsive page with an unsupported op (e.g. format_text, position_element, resize_element, insert_fill), do NOT make the call — tell the user that operation isn't supported on that page and offer an alternative.
The commit approval gate (required)
commit-editing-transaction makes changes permanent. You MUST show the user exactly what changed (and the preview thumbnail) and get explicit approval before committing — e.g. "Here's the preview. Save these changes to your design?" Wait for a clear yes.
- Do NOT commit without approval.
- Do NOT tell the user changes are saved before the commit call has succeeded.
- After a successful commit, give the user a direct link to open the design in Canva.
- If a commit fails, all changes are lost — start a new transaction to retry.
Note for composing skills: a skill that already collects a single up-front approval for a batch of changes (e.g.
canva-implement-feedback) should treat that approval as covering the commit and NOT ask again. Follow that skill's own confirmation rules; the gate above is the default for direct, ad-hoc edits.
Workflow
Step 1: Resolve the design
- Short link (
canva.link/...) →Canva:resolve-shortlinkto get the URL. - Full Canva URL → extract the design ID (the segment after
/design/). - Raw design ID (starts with
D) → use directly; do NOT search. - Nothing provided → ask for the design ID or link.
Step 2: Start the transaction and inspect
Call Canva:start-editing-transaction. Show the thumbnail(s). Use the returned content to locate the exact element_ids you need to target. (If you only needed to look, call cancel-editing-transaction and stop.)
Step 3: Build and perform operations
Translate the user's request into concrete operations. Confirm scope first when a find_and_replace_text string could match in multiple places or contexts — ask which instances they mean. Batch all operations into one perform-editing-operations call when you can.
Step 4: Preview and commit
Show the resulting thumbnail and a plain-language list of what changed. Ask for approval, then commit-editing-transaction. Share the edit link.
Rules
- Always remember and reuse the
transaction_idandpagesarray within a transaction. - Never leave a transaction uncommitted without telling the user their draft was discarded.
- For destructive ops (
delete_element, largefind_and_replace_text), confirm scope before performing. - Prefer one batched
perform-editing-operationscall over many small ones.
04 同仓库更多 Skills
skill-installer
Install Codex skills into $CODEX_HOME/skills from a curated list or a GitHub repo path. Use when a user asks to list installable skills, install a curated skill, or install a skill from another repo (including private repos).
skill-creator
Create or update a Codex skill with appropriately scoped instructions and any needed supporting resources.
review-agent
Perform a read-only, defect-first review of a specified code change and return every actionable finding. Use when another agent delegates review of uncommitted changes, a base-branch diff, a commit, or custom review instructions.
plugin-creator
Create and scaffold plugin directories for Codex with a required `.codex-plugin/plugin.json`, optional plugin folders/files, valid manifest defaults, and personal-marketplace entries by default. Use when Codex needs to create a new personal plugin, add optional plugin structure, generate or update marketplace entries for plugin ordering and availability metadata, or update an existing local plugin during development with the CLI-driven cachebuster and reinstall flow.
openai-docs
Use for Codex models/pricing, scheduled tasks, skills, settings, setup, troubleshooting, customization, automations, and self-knowledge—including 'you,' 'your,' 'this app,' or 'this coding agent' when they refer to Codex—and for OpenAI APIs/products and ChatGPT Work. Also use for model choice/migration, prompting, SDKs, Responses, Realtime, agents, evals, and Chat/Work/Codex comparisons. Do not use for generic app/software tasks that merely mention Codex.
imagegen
Generate or edit raster images when the task benefits from AI-created bitmap visuals such as photos, illustrations, textures, sprites, mockups, or transparent-background cutouts. Use when Codex should create a brand-new image, transform an existing image, or derive visual variants from references, and the output should be a bitmap asset rather than repo-native code or vector. Do not use when the task is better handled by editing existing SVG/vector/code-native assets, extending an established icon or logo system, or building the visual directly in HTML/CSS/canvas.