Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add anshrajore/atelier-mcp --skill ateliergit clone --depth 1 https://github.com/anshrajore/atelier-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/anshrajore/atelier-mcp/atelier)<a href="https://agentmods.dev/skills/anshrajore/atelier-mcp/atelier"><img src="https://agentmods.dev/badge/skills/anshrajore/atelier-mcp/atelier/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/anshrajore/atelier-mcp/atelier"><img src="https://agentmods.dev/badge/skills/anshrajore/atelier-mcp/atelier.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00034 | $0.01801 |
| Opus 5 | $0.00017 | $0.00901 |
| Sonnet 5 | $0.00007 | $0.00360 |
| Haiku 4.5 | $0.00003 | $0.00180 |
Grade A, and why
atelier scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 9d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- **check:** `ast: CallExpression(name in ['fetch', 'axios.get', 'axios.post', 'http.request']) -> if !has_property('timeout') and !has_property('signal') -> fail` How it starts
The opening of the file, as written. The whole thing — 96 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Atelier Canonical Ruleset (Base)
Atelier Base defines universal, gradeable quality gate rules that apply across any stack. Every rule contains an absolute pass/fail threshold and a mechanical check: specification for automated evaluation without subjective interpretation.
1. Universal UI/UX Rules (BASE-UI-xxx)
BASE-UI-101: Harmonic Spacing System Discipline
- Statement: All layout margins, paddings, and container gaps must strictly conform to an 8px harmonic grid (subgrid: 4px).
- Threshold: Zero occurrences of spacing values outside
[0, 4, 8, 12, 16, 24, 32, 48, 64, 96, 128]px. - Violation Example:
margin-top: 13px;orpadding: 7px 19px;orp-[17px]. - check:
regex: (?:(?:margin|padding|gap|top|bottom|left|right)(?:-(?:top|bottom|left|right|x|y))?\s*:\s*(\d+)px|(?:m|p|gap)-?\[(\d+)px\]) -> extract integer $1 or $2 -> fail if num % 4 != 0 or num not in [0,4,8,12,16,24,32,48,64,96,128]
BASE-UI-102: Strict Typographic Scale & Optical Tracking
- Statement: Font sizes must derive from a standardized modular scale, with negative tracking on display text (>=24px) and wide tracking on micro-labels (<=12px uppercase).
- Threshold: Font size values must belong to
{12, 14, 16, 18, 20, 24, 30, 36, 48, 60, 72}px. Display headers (>=24px) must haveletter-spacing <= -0.015em. - Violation Example:
font-size: 29px;orfont-size: 36px; letter-spacing: normal;. - check:
ast: CSS/JSX TextStyleNode -> where fontSize >= 24px and letterSpacing not in ['-0.02em', '-0.025em', '-0.03em', '-0.04em', 'tracking-tight', 'tracking-tighter'] -> fail
BASE-UI-103: WCAG 2.1 AA Contrast Minimum
- Statement: Body text and interactive icons must maintain a minimum contrast ratio of 4.5:1 against their rendered background (3.0:1 for text >= 24px).
- Threshold: Calculated relative luminance contrast ratio
< 4.5:1for regular text (<24px) or< 3.0:1for large text (>=24px) is a hard failure. - Violation Example: Text
#94a3b8(Slate-400) on white#ffffffbackground (contrast ratio: 2.45:1). - check:
metric: contrast_ratio(foreground_color, background_color) -> fail if (is_large_text ? ratio < 3.0 : ratio < 4.5)
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 9d ago First seen · 96 lines · 34 tokens per session scan A 10742b771098
atelier is a skill published in the GitHub repository anshrajore/atelier-mcp (0 stars, last pushed 22d ago), licensed MIT. It adds 34 tokens to every session and 1,801 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
interactive-ui
Ask the user to pick an option, or render GitHub data as an interactive UI panel.
pr-analysis
Analyse a GitHub pull request by fetching its metadata, diff, linked issues and CI status.
pr-review
Review a GitHub pull request by posting inline code comments and submitting a review decision.
issue-management
Create, update, list and search GitHub issues and PRs, list a repository's labels, run milestones, and check for duplicates.
release-management
Tag a commit, publish a GitHub release, and read, correct or withdraw the tags and releases already there.
error-handling
Read the error codes these tools raise and decide whether to retry, fix the call, or stop.