# UI Craft (npm · ui-craft-mcp)

UI quality checks: anti-slop, accessibility, design tokens, scoring and workflow routing.

- Trust score: 82/100 (high trust)
- Registry status: active
- Liveness: live
- Owner verified: no
- Last scored: 2026-11-21

## Components

- npm · `ui-craft-mcp`: 82/100 (this document), [markdown](https://verifymcp.io/servers/educlopez-ui-craft/ui-craft-mcp.md), [page](https://verifymcp.io/servers/educlopez-ui-craft/ui-craft-mcp)

## Channel facts

- Registry: `npm`
- Package: `ui-craft-mcp`
- Version: `0.9.1`
- Transport: `stdio`

## Trust breakdown

How this component scores in each security and reliability category. Every signal is checked automatically from public evidence about the published package, including repeated runs of it in an isolated sandbox, and we only credit what we can confirm. Scores are 0–100 per category. Scoring method: https://verifymcp.io/docs/scoring (what has changed: https://verifymcp.io/docs/scoring/changelog)

Scored 2026-11-21.

- **Supply Chain Security**: 98/100
  - No malware found by supply-chain analysis.
  - No known CVEs affecting this package version or its production dependencies.
  - No install/post-install scripts declared.
  - 31 of 92 dependencies flagged as unhealthy.
- **Provenance & Transparency**: 100/100
  - Source repository is publicly reachable at the declared URL.
  - Cryptographically verified build provenance (signed, bound to educlopez/ui-craft).
  - Clear OSI-approved license (MIT).
  - Actively maintained (last published 4 days ago).
  - Publishes a security disclosure policy (SECURITY.md).
- **Schema Quality & AI Usability**: 69/100
  - AI-judged instruction clarity (excellent).
  - Context-footprint check failed: tool/resource definitions use about 1442 tokens (~180/item across 8 items; 8 tools + 0 resources), over budget; trim descriptions and params.
  - Usage-examples check failed: none of the tools include examples.
- **Stability & Change Management**: 17/100
  - Stability observed for 5 of 30 days with no destabilising changes; credit accrues until the full window elapses.
- **Tool Coverage**: 100/100
  - 100% of tools have a non-trivial description (not blank, and not just the tool's name).
  - 100% of tool parameters carry a description.
  - Structured output schemas are declared (100% of tools); any adoption earns full credit.
- **Tool Safety**: 100/100
  - No prompt-injection markers were found in the server instructions, tool names or descriptions we captured.
  - We read all 8 captured tool definition(s), and no name or description among them implies an irreversible operation.
  - An AI judge read all 8 captured unit(s) of tool text and found none that tries to manipulate the model reading it.
- **Capabilities**: 100/100
  - Implements a supported MCP spec version (2025-11-25); the latest is 2026-07-28.

## Install

### How do I install the UI Craft MCP server?

UI Craft runs locally as an npm package, launched with npx -y ui-craft-mcp. Ready-made configuration for Claude, Cursor, VS Code, Codex and 5 more is on this page, copied from each client's own documentation.

### Claude

```bash
claude mcp add educlopez-ui-craft -- npx -y ui-craft-mcp
```

### Cursor

```json
{
  "mcpServers": {
    "educlopez-ui-craft": {
      "command": "npx",
      "args": [
        "-y",
        "ui-craft-mcp"
      ]
    }
  }
}
```

### VS Code

```json
{
  "servers": {
    "educlopez-ui-craft": {
      "command": "npx",
      "args": [
        "-y",
        "ui-craft-mcp"
      ]
    }
  }
}
```

### Codex

```bash
codex mcp add educlopez-ui-craft -- npx -y ui-craft-mcp
```

### opencode

```json
{
  "$schema": "https://opencode.ai/config.json",
  "mcp": {
    "educlopez-ui-craft": {
      "type": "local",
      "command": [
        "npx",
        "-y",
        "ui-craft-mcp"
      ],
      "enabled": true
    }
  }
}
```

### OpenClaw

```bash
openclaw mcp add educlopez-ui-craft --command npx --arg -y --arg ui-craft-mcp
```

### Hermes

```yaml
mcp_servers:
  educlopez-ui-craft:
    command: "npx"
    args: ["-y", "ui-craft-mcp"]
```

### Netclaw

```json
{
  "McpServers": {
    "educlopez-ui-craft": {
      "Transport": "stdio",
      "Command": "npx",
      "Arguments": [
        "-y",
        "ui-craft-mcp"
      ]
    }
  }
}
```

### Vellum

```bash
assistant mcp add educlopez-ui-craft -t stdio -c npx -a -y ui-craft-mcp
```

### Other

```json
{
  "mcpServers": {
    "educlopez-ui-craft": {
      "command": "npx",
      "args": [
        "-y",
        "ui-craft-mcp"
      ]
    }
  }
}
```

## Changelog

Every change recorded for this component, newest first. Days that predate change tracking, or that we cannot explain, say so: "we were watching and nothing happened" and "we were not watching" are different claims.

### 2026-11-20 (score 82, +1)

No change was recorded against any check on this day. Stability & Change Management went from 10 to 13. That category is still filling its 30-day observation window: 3 days of observed history at the previous scan, 4 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-18 (score 81, +1)

No change was recorded against any check on this day. Stability & Change Management went from 3 to 7. That category is still filling its 30-day observation window: 1 days of observed history at the previous scan, 2 at this one. The score rises as the window fills, whether or not the server changes.

### 2026-11-16 (score 80)

First indexed and scored.

## MCP tools (8)

### `route_task` (~184 tokens)

Route Task

Call this FIRST on any UI request, before reading reference files. Takes the task in natural language and returns the references, commands and MCP tools that cover it, ranked, plus the single first move to make. Deterministic lexical ranking over a synonym map, so a prompt reaches the right entry even when its words do not match our filenames — "analytics panel" finds recipe-dashboard.md, which never says analytics. Recipes are also indexed by what they contain, so "pricing block" finds recipe-landing.md. Returns POINTERS only: every design rule stays in the files it points at. An unmatched prompt is reported as unmatched, not routed to a guess.

Input parameters:

- `limit` (integer): Max hits per kind (default 5, clamped to 12)
- `prompt` (string, required): The task in natural language — e.g. "analytics panel with KPIs"

Output parameters:

- `concepts` (array)
- `error` (string)
- `first_move` (string|null)
- `instruction` (string)
- `intent` (string)
- `prompt` (string|null)
- `results` (object)

### `check_anti_slop` (~117 tokens)

Check Anti-Slop

Scans source code for anti-slop violations using the deterministic rules from ui-craft-detect. Accepts either a `code` string (inline source) or a `path` string (file or directory). Returns findings with severity, rule ID, file, line, and message. These are the 43 deterministic rules only — no taste or aesthetic judgment.

Input parameters:

- `code` (string): Inline source code to scan (alternative to path)
- `path` (string): File or directory path to scan (alternative to code)

Output parameters:

- `coverage`
- `error` (string)
- `findings` (array)
- `scan_errors` (array)
- `scan_policy`
- `summary` (object)
- `version` (string)

### `tokens_lint` (~101 tokens)

Tokens Lint

Static analysis for off-system token values: raw hex colors, non-scale border-radius px values, non-8pt spacing values, and magic z-index integers. Token scale source of truth: references/tokens.md. Accepts `code` string or `path`. Returns structured findings + summary.

Input parameters:

- `code` (string): Inline source code to lint (alternative to path)
- `path` (string): File or directory path to lint (alternative to code)

Output parameters:

- `coverage`
- `error` (string)
- `findings` (array)
- `scan_errors` (array)
- `scan_policy`
- `summary` (object)
- `version` (string)

### `acceptance_bar` (~90 tokens)

Acceptance Bar

Returns the deterministic acceptance checklist for a UI surface. Data is bundled from recipe-dashboard.md, recipe-landing.md, recipe-auth.md, and finish-bar.md. Surfaces: dashboard, landing, auth, generic. Returns DATA only — no scoring or judgment. Scoring uses check_anti_slop + tokens_lint results.

Input parameters:

- `surface` (string, required): The UI surface to retrieve the acceptance bar for

Output parameters:

- `error` (string)
- `items` (array)
- `surface` (string|null)

### `ux_coverage` (~214 tokens)

UX Coverage

Returns the parts a screen of a given kind needs in order to be complete, and the contract for reporting them. This is the COMPLETENESS axis — "does this screen have the parts screens of its kind need". It is not the distinction axis: acceptance_bar and score_ui answer "is this designed". Report the two side by side; never fold coverage into a score, a count or a percentage. Report-only — coverage never gates and never fails a build. Archetypes: data-table, settings, search-filter, detail-view, empty-first-run, billing, pricing, docs-page, checkout, onboarding, destructive-confirm, invite-share. Call with no archetype to get the catalogue. dashboard, landing and auth are served by acceptance_bar instead. Returns DATA only — no scoring or judgment.

Input parameters:

- `archetype` (string): Screen archetype to retrieve coverage parts for. Accepts the key, the label, or a synonym ("table", "plans", "are you sure"). Omit to list the catalogue.

Output parameters:

- `archetype` (string|null)
- `catalogue` (array)
- `error` (string)
- `family` (string)
- `items` (array)
- `label` (string)
- `note` (string)
- `reporting` (object)
- `states` (array)

### `score_ui` (~255 tokens)

Score UI

Composite design-quality scorer (UICraftScore). Combines three deterministic dimensions into a single 0-100 score + letter grade (A/B/C/D/F) + per-dimension subscores and findings. Dimensions: anti-slop (38 rules via ui-craft-detect), token-discipline (raw hex / off-scale values), and static a11y (5 checks: img-no-alt, non-semantic-interactive, positive-tabindex, aria-invalid-no-describedby, no-reduced-motion). Accepts either a `code` string (inline source) or a `path` string (file). Formula: score = 100 − (antiSlop_crit×8) − (antiSlop_major×4) − (antiSlop_warn×1) − (token_findings×2) − (a11y_crit×8) − (a11y_major×4), clamped [0,100]. Returns { overall: {score, grade}, dimensions: {anti_slop, token_discipline, a11y}, version }.

Input parameters:

- `code` (string): Inline source code to score (alternative to path)
- `path` (string): File path to score (alternative to code)

Output parameters:

- `coverage`
- `dimensions` (object)
- `error` (string)
- `overall` (object)
- `scan_errors` (array)
- `scan_policy`
- `version` (string)

### `fold_candidates` (~220 tokens)

Fold Candidates

Draws composition classes for a landing fold, preferring ones this project has not spent yet. Variety cannot be requested in prose — asking a model to "be different" returns its default every time — so the class is drawn instead — ordered by a per-project seed, then by what this project has already spent. Pass the classes already used in this project as `used`. Classes: type-only, full-bleed-overlay, split, stacked, product-dominant, band. `split` is drawn last on purpose: it is the fold every generator reaches for unprompted. Returns each class with what it demands and what it sacrifices — the sacrifice is what makes a composition distinct.

Input parameters:

- `count` (integer): How many candidates to draw (default 3)
- `seed` (string): Overrides the per-project seed. Defaults to the working directory, which is what makes two projects draw differently while one project stays stable across runs.
- `used` (array): Composition class ids already used in this project (deprioritised in the draw)

Output parameters:

- `candidates` (array)
- `error` (string)
- `instruction` (string)
- `seeded_by`: What the order was seeded by, so a caller knows it will repeat. Deliberately not a value derived from the seed: a digest of a guessable path is not redaction.
- `used` (array)

### `check_fold` (~261 tokens)

Check Fold

Renders a URL and reports what its fold actually is: a screenshot, the composition class it belongs to, drift from the class you set out to build, three judged invariants, and four measurements reported WITHOUT a verdict. The unjudged four (identification, single dominance, asymmetry, restraint budget) are measured but not scored: calibration against reference landing pages showed those thresholds do not yet discriminate, so they are yours to interpret. Treat the screenshot as the primary output — look at it. Invariant 7 (a costly detail) is unmeasurable by construction: pass what you committed to as `costly_detail`, or it fails. Needs a browser: install puppeteer, or set UI_CRAFT_CHROME to an existing Chrome to skip the download.

Input parameters:

- `costly_detail` (string): The one element of this fold that could not have come from a template, and what it cost
- `expected_class` (string): The composition class you set out to build, to detect drift
- `height` (integer): Viewport height (default 900)
- `url` (string, required): URL of the rendered page — a local dev server URL is fine
- `width` (integer): Viewport width (default 1440)

Output parameters:

- `checks` (array)
- `composition` (object)
- `drift` (string|null)
- `drift_note` (string): Why nothing was compared, when that is the case
- `drift_status` (string): Whether the built class was compared against an intended one at all. 'not-compared' means no expected_class was passed — an absent drift field is not evidence of none.
- `error` (string)
- `expected_class` (string|null)
- `measured`
- `observations` (array)
- `summary` (object)
- `url` (string)

## Diagnostics

Captured diagnostic sections: Provenance, Dependencies. The full working is on the page: https://verifymcp.io/servers/educlopez-ui-craft/ui-craft-mcp#diagnostics

## Score history

- 2026-11-21: 82
- 2026-11-20: 82
- 2026-11-19: 81
- 2026-11-18: 81
- 2026-11-17: 80
- 2026-11-16: 80

## Common questions

### What is the UI Craft MCP server?

UI Craft is an MCP server listed in the public MCP registry as io.github.educlopez/ui-craft. UI quality checks: anti-slop, accessibility, design tokens, scoring and workflow routing. This page covers its npm package (ui-craft-mcp).

### Is the UI Craft MCP server safe to use?

UI Craft scores 82 out of 100 on VerifyMCP. We found no known CVEs affecting it as of 21 November 2026. It declares no install or post-install scripts. Its build provenance is signed and verified. That is a record of what we were able to check automatically, not an endorsement. The category breakdown on this page shows every signal behind the number, including the ones we could not confirm.

### What tools does the UI Craft MCP server expose?

UI Craft exposes 8 tools: route_task, check_anti_slop, tokens_lint, acceptance_bar, ux_coverage, and 3 more. Their descriptions and schemas cost roughly 1,442 tokens of context every time the server is loaded.

### Is the UI Craft MCP server still maintained?

UI Craft is still listed as active in the MCP registry. We last reached this channel on 21 November 2026. Those dates come from our own scans of the registry and the channel itself, not from anything the publisher announced.

### What licence is the UI Craft MCP server under?

UI Craft declares the MIT licence, which is OSI-approved. That covers the source only, and says nothing about the cost of any service it calls.

### What did VerifyMCP test on the UI Craft MCP server?

VerifyMCP ran 16 checks on UI Craft, covering Malware scan, Known CVEs, Install scripts, Dependency health and 12 more. Endpoint Security and Transport & Reachability do not apply to an npm package, so they were not tested. The score does not measure everything, and the scoring documentation lists what it leaves out.

## Links

- npm package: https://www.npmjs.com/package/ui-craft-mcp
- Socket report: https://socket.dev/npm/package/ui-craft-mcp
- Repository: https://github.com/educlopez/ui-craft
- Changelog RSS feed: https://verifymcp.io/servers/educlopez-ui-craft/ui-craft-mcp.xml
- Changelog JSON feed: https://verifymcp.io/servers/educlopez-ui-craft/ui-craft-mcp/changelog.json
- Trust report (JSON): https://verifymcp.io/servers/educlopez-ui-craft/ui-craft-mcp.json
- HTML version of this page: https://verifymcp.io/servers/educlopez-ui-craft/ui-craft-mcp
