Skip to main content
Glama

Minerals

Find design libraries and tools

find_design
Read-onlyIdempotent

Searches the Minerals shelf of design libraries and tools. Applies when your person asks for, or you are about to choose, a UI component, block, texture, animation, icon set, typeface, colour system, sound or design reference, or asks to change one. Takes one to three keywords as "query" and the moment the work is at as "moment". Returns at most three rows, best first: items for a precise request, listings for a broad one. With nothing sent, it returns the six moments and what is on the shelf. What it returns is a suggestion for your person. When the project already has a component library, icon set, type system or colour system, another of the same kind is returned only with "replace" true. Refs sent as "have" or "declined" from minerals.json are left out, and with "quiet" true nothing is suggested.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "find_design"
  }
}

Input Schema

NameRequiredDescriptionDefault
haveNoThe refs already in the project's minerals.json. Read that file and send them on every request: they are never returned again, and what goes with them ranks higher.
kindNoLimit to one kind: components, blocks, textures, motion, illustration, icons, type, color, sound, canvas, reference, skills, generators, people.
needNoOlder name for query. Still accepted and treated as the query.
queryNoOne to three keywords for the thing wanted, e.g. "tooltip", "pricing" or "share review".
quietNo"quiet" from minerals.json. When true, nothing is suggested; search still works.
stackNoLimit to supply that works in this stack, e.g. "React" or "nextjs, tailwind". Several names may be given; any one matching is enough.
stageNoThe project's stage, one of: new, existing-system, polish. Worked out and stored the same way as project.
momentNoWhere the work is right now, one of: Starting, Building a section, Filling a gap, Polishing, Tuning, Sharing. Pick it from what your person said. It decides which listings may answer.
listingNoLimit to one listing by its ref. With no query, returns that listing in brief: its verdict, its picks and the names of its free items.
projectNoThe project type, one of: marketing-site, product-app, dashboard, brand, shop, docs, portfolio, mobile-app. Work it out yourself from the project's routes, dependencies and pages; do not ask your person. Send only the label. Store it in minerals.json under "project" as {"type", "stage"} and read it from there on later calls; your person can edit that file.
replaceNoSend true only when your person asked to change or replace what the project has, or named another library.
sectionNoLimit to one section: libraries or tools.
acceptedNoListings your person has approved once for this project, from minerals.json. An item from one of them needs no new yes.
declinedNoThe "declined" list in the project's minerals.json: what the person said "not now" to. It is left out.
free_onlyNoLeave out supply that cannot be used without paying.

Schema Changelog

Changes observed during successful MCP inspections.

  1. First observed

TDQS

A4.2/5.0
Behavior5/5

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

Annotations only cover the read-only/idempotent safety profile, while the description adds substantial behavioral context: result cap of three, best-first ordering, empty-argument behavior returning the six moments and shelf contents, the replace gate for existing systems, exclusion of have/declined refs, and suppression via quiet. It also frames output as a suggestion rather than an applied change, which is critical for an agent deciding whether to act.

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?

Purpose and trigger are front-loaded, then return behavior, then gating rules — a sensible ordering. It is dense but nearly every sentence carries a distinct rule; only the 'What it returns is a suggestion for your person' framing mildly overlaps with the preceding return-shape sentence.

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?

With 15 optional parameters, no output schema, and only safety annotations, the description does the heavy lifting on return shape (max three rows, ordering, empty-call fallback) and on the minerals.json persistence contract. It is adequate for correct invocation, though it doesn't tie together how project/stage/stack narrow results alongside moment.

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 100%, so baseline is 3, but the description adds real meaning about interactions: query takes one to three keywords, moment selects which listings may answer, have refs are never returned and boost ranking of neighbors, and replace gates same-kind results. The remaining nine parameters (kind, stack, stage, project, listing, section, accepted, free_only, need) are left entirely to the schema.

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?

States a specific verb and resource — 'Searches the Minerals shelf of design libraries and tools' — and enumerates the artifact types it covers (component, block, texture, icon set, typeface, colour system). The return shape ('at most three rows, best first: items for a precise request, listings for a broad one') distinguishes it from a single-listing fetch like get_listing, but no sibling is named explicitly.

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?

Trigger conditions are unusually explicit: use it when the person asks for an artifact, when the agent is about to choose one, or when they ask to change one. It also states when-not conditions indirectly ('another of the same kind is returned only with replace true', 'with quiet true nothing is suggested'). What it lacks is routing to alternatives such as get_listing or suggest_listing.

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

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.