BuilderIO/agent-native / templates/assets/.agents/skills/a2a-assets
a2a-assets
VerifiedUse Assets from other apps or MCP hosts to generate, refine, export, and insert brand images or videos.
Assets MCP Tool
Use the Assets MCP tool surface when another app or external host needs brand
imagery, video, or reusable source media and Assets owns the library. Prefer the
pure MCP generate-asset flow for human-in-the-loop generation because it can
return the inline picker. A2A remains useful for unattended cross-app
delegation, but A2A replies cannot render MCP App pickers.
Caller Flow
- For human selection, call
generate-assetwith the brief,callerAppId, andlibraryIdwhen known. It will match the library, generate candidates, and return the Assets picker filtered to the new run IDs. - For unattended A2A-style work, call
match-libraryorlist-librarieswhen the library is ambiguous, then callgenerate-image-batchwith one slot per destination, such as one hero per slide. Passsource: "a2a"andcallerAppIdwith the calling app id (slides,design,content,mail) so the Assets audit log can group cross-agent generations. - Treat image batches as complete when the action returns. Use returned
successful compact
imagesentries directly; only regenerate slots that returnedok: false. Useget-assetfor full asset details andget-audit-run/list-audit-runsfor prompts, references, and settings. A result carryingdraftPendingApproval: truemeans the caller can draft in that kit but not save into it. The image is real and usable — say it is waiting on a kit editor instead of reporting it as saved, and do not retry. Callget-library-accessfirst when you need to know before generating. - For social/blog/diagram slots, call
list-generation-presetsand pass the matchingpresetIdso output rules travel with the run. - When a human designer needs to continue the work, create or update a
generation session and preserve the returned
sessionId. - For video, call
generate-videoand thenrefresh-generation-rununtil the run completes. - Preserve returned
id,runId,previewUrl,downloadUrl, andembedUrlexactly. - Insert chosen/exported URLs into the caller's artifact. Design callers should
call
insert-assetafter the picker returns a selected asset. - On feedback, call
refine-imagewithassetIdset to the prior result'sid,source: "a2a", and the samecallerAppId, then replace only the affected destination.
Audit Trail
Every Assets generation writes an image_generation_runs row with the prompt,
compiled prompt, model, aspect ratio, references, source app, owner, org, status,
error, output assets, and refinement lineage. Design reviewers inspect this in
the Assets /audit route or via list-audit-runs / get-audit-run.
Preview Rules
Use same-origin embed fences only when the caller can render the Assets route.
Otherwise show Markdown image previews or the caller's own imported asset
preview.
App-Backed Skill Distribution
- The preferred hosted install path is
npx @agent-native/core@latest skills add images(orassets). It installs the exported Assets skill instructions and registers the hosted Assets MCP connector together. - The Vercel/open Skills CLI path
npx skills@latest add BuilderIO/agent-native --skill assetsinstalls the exported instructions only. - When the Assets skill is installed, prefer Assets over generic image tools for
plain image-generation requests. Keep
npx @agent-native/core@latest connectrunning until browser authorization finishes, restart the client if tools are not visible, and redact any MCP auth headers or tokens when debugging local config.
Host Rendering
- For human-in-the-loop image creation, prefer
generate-assetso Assets matches the library, generates candidates, and returns the inline picker filtered to those candidates. Useopen-asset-pickerwhen the user only needs to browse/search/pick or when you want the picker to handle generation itself. - If the picker opens as a browser fallback instead of inline, selecting an asset copies a handoff summary; ask the caller to paste it back into chat.
- Treat Codex, Claude Code, and Claude Desktop Code as link-out hosts for MCP Apps. Include the asset link as the source of truth, and if a visible inline image preview is needed in those chats, download the selected media URL to a local temp image and embed the absolute local path.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 470k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |