Name: Atris
Author: atrislabs

Atris MCP

Run Atris campaigns from any agent, and extract, check, or search brand design systems.

Grade B Local · stdio Optional API key 10 tools v3.80.1 · 6 Nov 2026 MIT

No reviews yet by atrislabsnpm 46,925 Runs on your machine, nothing hosted

83/100

trust score

downloads a day

local

runs on your machine

14k

downloads in 30d

What it does

Run Atris campaigns from any agent, and extract, check, or search brand design systems.

Quickstart

# 1 — add it to Claude Code
claude mcp add --scope user atris-mcp '--' npx -y atris mcp

# 2 — ask your agent for something these tools do
#   design_extract, design_check, design_search and 7 more

Atris costs nothing on mcprush: there is no plan to choose here, no cap to set and nothing mcprush can bill you for.

Publishernot claimed

atrislabs1 server

1 category

All 1 server

Collected from a public index. Nobody has claimed this account, so nothing here was written by its author — claim it if it is yours.

Installation

Where are you running it?

Every route below installs the same thing and ends at the same approval screen. This one runs on your machine: your client starts Atris as a process under your own user, with your files and your network, so the tool surface below is what it can do to your computer rather than to a server somewhere else. Its scan result is on the Security tab — read the surface before you approve it.

Claude Codedeclared by this listing

This is a public server: you run it yourself and this marketplace is not in the path. You need Node.js — the current LTS (24) is the safe choice; npx comes with it. Claude Code registers it in one command; --scope user makes it available in every project. The '--' is in quotes so that PowerShell passes it on.

claude mcp add --scope user atris-mcp '--' npx -y atris mcp

Reconnect, or start a new session, and the tools appear in the model’s tool list.

Free on mcprush, and it runs from its publisher’s own package: this marketplace is not in the path, so there is no key of ours to hold and no card to add — some of its tools take a key of its publisher’s own, and the rest work without.

What it asks of you

Nothing is required, but it can take ATRIS_API_KEY (Atris API key for the design tools. Campaign tools work without it.) if you want to set it. Its publisher says some of its tools need it, and the rest work without.

What it writes

One config entry your client uses to start the process locally. A local server runs with your file system and your network, which is why it is priced without metering.

What you approve

10 tools, with what each one reads, writes and reaches shown before you agree — the same list on every route above. Read the tool surface.

Tool surface

What the model actually sees. Descriptions are diffed on every release — see version history.

Only tools that write or reach the network

Read from its source code. A server that builds its tool list when it starts can list more tools than this.

extract a site's design system (colors, typography, layout, voice) as json. costs 10 credits, 2 on a cache hit.

Parameter	Type	Required	Notes
url	string	yes	site url to extract, e.g. https://stripe.com

score how closely a page follows a reference brand. costs 20 credits.

Parameter	Type	Required	Notes
source_url	string	yes	page being checked
reference_url	string	yes	brand url to check against

search brands already extracted by atris design. costs 1 credit.

Parameter	Type	Required	Notes
query	string	yes	plain words, e.g. "dark developer tools brand"
limit	number	no	max results

list every atris campaign with its state, money spent, and the next move. a campaign is a long goal with a judge (a command that exits 0 when the goal is met), a budget, and members.

Parameter	Type	Required	Notes
cwd		no	—

show one campaign: judge and its last reading, budget, seats, missions, open asks, and the next move.

Parameter	Type	Required	Notes
id		yes	—
cwd		no	—

the easiest start: give a goal in plain words. the owner's engine drafts a judge script, atris runs it to prove it fails today, and returns the draft. start: true also starts the campaign. budget defaults to $20.

Parameter	Type	Required	Notes
goal	string	yes	the goal in plain words
budget	number	no	dollars the campaign may spend (default 20)
owner	string	no	member who owns it (default mission-lead)
start	boolean	no	start the campaign once the drafted judge fails today
cwd		no	—

start a campaign with a judge you wrote. the judge runs first and a judge that already passes is refused: it must fail today. print graded output like "ℹ pass 3" and "ℹ fail 2" so progress shows.

Parameter	Type	Required	Notes
title	string	yes	short name for the goal
judge	string	yes	shell command that exits 0 only when the goal is met
budget	number	yes	dollars the campaign may spend
owner	string	no	member who owns it (default mission-lead)
seats	array	no	members who work on it, as member or member:role
bet	string	no	what you think will work
id	string	no	optional id, cmp-<lowercase-words>
as		no	—
cwd		no	—

let the campaign run itself for at most max_steps steps: it judges, proposes missions, and runs them. it stops early when the judge passes, the money runs out, or a person must answer an ask. returns the stop reason.

Parameter	Type	Required	Notes
id		yes	—
max_steps	number	no	steps before it stops, 1 to 20 (default 3)
parallel	number	no	missions at once, 1 to 5 (default 3)
as		no	—
cwd		no	—

stop for a person before anything that cannot be undone: sending, paying, deleting, deploying, publishing. the campaign waits until a person answers.

Parameter	Type	Required	Notes
id		yes	—
question	string	yes	the question, with what happens on yes
mission	string	no	mission id the ask belongs to
as		no	—
cwd		no	—

pass along a person's answer to an open ask. an agent cannot close an ask: the answer is recorded and the ask stays open until the person runs the returned person_command in their own terminal.

Parameter	Type	Required	Notes
id		yes	—
ask_id	string	yes	ask id from campaign_show, e.g. a-1b2c3d
answer	string	yes	the person's answer in their words
as		no	—
cwd		no	—

Free

$0 forever

Every tool, no call limit
No card, and mcprush charges nothing for it
Source published under MIT
Runs on your machine — nothing of it reaches our gateway
Nothing to cap, because mcprush bills nothing

Nothing here is billable: mcprush charges nothing to install Atris or to call it, at any volume.

BTrust score 83 / 100

Release 3.80.1 · scanned 6 Nov 2026

Two independent axes, because powerful and malicious are different questions. The grade is threat only. The capability level is blast radius, and it is never a penalty on the grade — it is priced as one subtract-only term in the score, where you can see it.

Threat score89

Client score83

Findings18 threat · 0 capability

Capabilityhigh

Coveragesource

Verificationsource

Surface digestc7e67b65c6fa4e2f5ae400fce6624cf9a3d3f68c036237d49977e6db8044d8e5

Score vectorthreat 89 → client 83

Term	Level	What it prices	Points
capability-exposure	high	capability blast radius (high) — client exposure if the model is manipulated	−6
verification-discount	source	publisher verification (provenance) — cryptographic build provenance ties the artifact to its source	−0
coverage-honesty	source	inspection depth (source) — how much of the target the scan could see	−0

score = clamp(round(threatScore − E_cap − E_ver − E_cov), 0, 100). No model, no opinion, no network call at scoring time: the same methodology version against the same bytes always produces this number, and every line above is a field of the report rather than a summary of it.

What the scan could actually read

A grade is only as meaningful as its coverage, so the scanner publishes its own depth before it publishes its result.

Tool surfacenot available

Implementation sourceread

Package metadataread

Live transportnot available

No tools were enumerated, so prompt-injection, capability and toxic-flow analysis had no tool surface to inspect. To grade a package’s real runtime tools, scan the running server: --command "npx -y <package>".

lib/task-prune.js (17 KB) was not read: over the file-count budget.

Capability — what it could do if the model were manipulated

Tags derived from each tool’s schema and the implementation, not from what the tool calls itself. high is the level these add up to.

Tool	Capability tags	Why the tag was assigned
No per-tool rows for this release — see the note below.

reads sensitive or local data

Toxic-flow graph

The lethal trifecta, checked as a graph rather than as a checklist: untrusted input, a sensitive source and an external sink have to meet before there is a path worth worrying about.

The public result for this release does not print the flow graph, so there is nothing to show here. That is not the same as "no paths were found": what the scan did read is above, under coverage.

Supply chain and provenance

Integrityfirst-seen

Surface digestc7e67b65c6fa4e2f

Scanned2026-11-06

This is the first scan of this surface here, so there is nothing yet to compare it against.

Every result on this tab comes from one deterministic pass over the published package — offline, rule by rule, and auditable line by line above. Same methodology version, same bytes, same score.