yardstick
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
No arguments | |||
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": false
} |
| prompts | {
"listChanged": false
} |
| resources | {
"subscribe": false,
"listChanged": false
} |
| experimental | {} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| searchC | Find stored references and seeded domain pointers by text, e.g. 'dark serif fintech landing'. |
| discoverA | Find reference sites for what the user is building, e.g. 'neighborhood dinner club landing page' or 'fintech dashboard for accountants'. Providers run in order (local catalog, live web search, then any configured command such as a subagent); one entry per site with a short why. Pass the domains to critique as --vs. |
| showC | Measurement summary and hero screenshot of a stored reference. |
| measureB | Measure any URL (localhost works) without storing it: fonts, type scale, spacing grid, palette, contrast, hero, layout. |
| critiqueA | Critique a URL against comma-separated reference names or domains (domains are captured on the fly). Returns a side-by-side table, findings with standard/gap/fix/evidence/region, the first change to make, and hero images (yours first). |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 5 tools
Search and discover both return reference sites, but search targets stored items via text while discover finds new sites from a project description, so the intent is mostly clear. Show, measure, and critique are each distinct and unambiguous.
All five tool names are single lowercase verbs following a consistent action-first pattern (search, discover, show, measure, critique), which is predictable and easy to follow.
Five tools is well-scoped for a design reference, measurement, and critique server. Each tool represents a meaningful, distinct action without excess or redundancy.
The server covers searching, measuring, and critiquing reference sites, but there is no tool to store or manage new references, which leaves a notable lifecycle gap. Agents can work around it by using domains on the fly, but the stored-reference workflow feels incomplete.