recipebooq

Propose an app design

propose_app

Your agent can design an app with you: send an app in opointo's design format (the AppSpec emit_app builds from) and get back a link that opens it in the App Builder as a new, unsaved design. The person sees every screen and the navigation, changes what they like, and hands it back for emit_app. Each problem comes back with its path and a fix. dryRun checks without making a link; format: true returns the format and every block.

Input Schema

TableJSON Schema

Name	Required	Description
`app`	No	The design: { name, tabs: [{ screenId, label }], screens: [{ id, name, composition: { screen: { content: { children: [node] }, chrome: { header?, headerItems?, floating: [] }, overlays: [] } } }], onboardingScreenIds: [], pushed: [{ screenId, fromScreenId, presentation }], theme? }. A node is { id, slug, props }: slug is "wireframe:<block>" such as "wireframe:list-row" (a list whose rows open one screen is one row block with repeats: "true" and its sample rows in rows: [{ title, ... }]) or a component from find_component, and a content node opens a screen with navigate: { to, presentation } beside its props. A top bar is header: { id, slug: "header", props: {} }, titled by the screen's name; a sheet has none. entryScreenId is only for a splash screen before the tabs. Tab icons, titles and composition metadata are set by the builder.
`dryRun`	No	check the design and report, without making a link
`format`	No

Schema Changelog

Changes observed during successful MCP inspections.

Changed3 schema fields changedOct 2, 2026

changedInput schema / properties / app / description

Previous value: -"The design: { name, tabs: [{ screenId, label }], screens: [{ id, name, composition: { screen: { content: { children: [node] }, chrome: { header?, headerItems?, floating: [] }, overlays: [] } } }], onboardingScreenIds: [], pushed: [{ screenId, fromScreenId, presentation }], theme? }. A node is { id, slug, props }: slug is a component from list_components or \"wireframe:<block>\" such as \"wireframe:list-row\" (a list whose rows open one screen is one row with repeats: \"true\"), and a content node opens a screen with navigate: { to, presentation }. A top bar is header: { id, slug: \"header\", props: {} }, titled by the screen's name. entryScreenId is only for a splash screen before the tabs. Tab icons, titles and composition metadata are set by the builder."New value: +"The design: { name, tabs: [{ screenId, label }], screens: [{ id, name, composition: { screen: { content: { children: [node] }, chrome: { header?, headerItems?, floating: [] }, overlays: [] } } }], onboardingScreenIds: [], pushed: [{ screenId, fromScreenId, presentation }], theme? }. A node is { id, slug, props }: slug is \"wireframe:<block>\" such as \"wireframe:list-row\" (a list whose rows open one screen is one row block with repeats: \"true\" and its sample rows in rows: [{ title, ... }]) or a component from find_component, and a content node opens a screen with navigate: { to, presentation } beside its props. A top bar is header: { id, slug: \"header\", props: {} }, titled by the screen's name; a sheet has none. entryScreenId is only for a splash screen before the tabs. Tab icons, titles and composition metadata are set by the builder."

addedInput schema / properties / format
```
Added value: +{
+  "type": "boolean"
+}
```
removedInput schema / required
```
Removed value: -[
-  "app"
-]
```

AddedSep 29, 2026

TDQS

A4/5.0

Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations mark it non-read-only, non-destructive and non-idempotent, and the description usefully adds that the result is an unsaved design (no persistence), that validation errors come back with a path and a fix, and that dryRun/format flags alter behavior. This is genuine behavioral context beyond the annotation flags.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the core action and the link result, then the dryRun/format caveats. The narrative middle ('The person sees every screen ... hands it back') is somewhat verbose but conveys the intended human-in-the-loop workflow, so it earns most of its space.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a complex nested-schema tool with no output schema, the description adequately explains what comes back (a link, or per-problem paths and fixes) and the flag-driven modes. A mention of the missing output shape or error/return specifics for the success path would round it out.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 67% and the 'format' parameter has no schema description, yet the description documents it ('returns the format and every block'), and dryRun is clarified as 'check ... without making a link'. The bulky app schema is mostly self-documenting, so description adds value where the schema is silent.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

The description gives a specific verb+resource ('design an app', 'send an app ... get back a link') and names the counterpart tool emit_app, so an agent can tell it apart from the emit path. It does not differentiate from the sibling propose_screen, which shares the 'propose' naming, leaving some ambiguity between the app-level and screen-level variants.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Usage context is clear: send a design to preview it as an unsaved App Builder link, then hand it back to emit_app. It also explains the dryRun branch ('checks without making a link'). No explicit when-not-to-use or alternatives beyond emit_app are stated.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All recipebooq tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources

Need Help?