MCP.Market

Umbra MCP server

by elberacasa·io.github.elberacasa/umbra·v1.2.0
No reviews yet

Trust score for AI-generated code: scan repos, guard agent file writes, get a 0-100 score.

B80/100grade B
Write a reviewInstall
What users say
No reviews yet
Be the first
Safety scan
B80/100

full report

Adoption
New

0 stars26 downloads/wk

Reviews

Write one

Nobody has reviewed Umbra yet.

If you have run it, two minutes of your experience saves the next person an afternoon.

Umbra tools (3)

write = sends, deletes, buys or posts

Read from the package source without running it. The installed server may list more.

  • get_score

    Fast static Umbra Trust Score (0-100) for a repository — SAFE and CLEAN axes only, no network, no sandbox. Use scan_repo for the full report with findings.

  • guard_content

    Check a proposed file write against the Umbra guard engine BEFORE writing. Returns the guard verdict as JSON: decision (allow/warn/block), findings with file:line evidence, and pathViolation when a protected path (e.g. .git/hooks) is targeted. On block, do not write — fix the content and re-check.

  • scan_repo

    Scan a repository and return the full Umbra Trust Score report as JSON (SAFE/CLEAN static findings; deep=true also verifies RUNS and HONEST in a Docker sandbox). Call before declaring a task done — a score below 50 means the work is not done.

Public scan report

scanner v0.1.9 · 2026-09-24 · same rubric, same numbers if you re-run it

1 medium
  • Code scan87 source files scanned20/25
  • –Live reliabilityno gateway calls yet and no remote to proben/a
  • –Tool poisoningtools not inspected (local package is not executed); not countedn/a
  • Auth qualitylocal package, no credentials required12/15
  • Maintenancelast push 50 days ago12/15
  • Maintainer identityregistry namespace matches repository owner; GitHub account older than a year8/10

Findings (1)

  • mediumeval / new Function usedexec.eval
    dist/guard/hook.js: …tead of string interpolation, and never eval() or new Function() input', 'safe/jw…
Overall 80/100. Components that don't apply are left out of the denominator. Any critical finding is an F.RubricAppeal a findingJSON

How Umbra compares

Against every graded AI and agents MCP server, from our own scans and uptime checks. Updated after each scan run.

Safety score
80/100
higher than 60% of 4,460 AI and agents servers
Tools
3
the typical AI and agents server has 6
Tools that change things
0%
typical: 0%. Fewer write tools means less an agent can break
Downloads a week
26
more than 12% of 1,538 AI and agents servers with a package

What its tools let an agent do: 3 read

Install Umbra in Claude Code, Cursor or VS Code

Runs npx -y @elberacasa/umbra on your machine. Read the scan report first; the gateway never runs local packages.

claude mcp add umbra -- npx -y @elberacasa/umbra
Add to Cursor

What the publisher says

From the Umbra repository's README, as published. We do not edit it. Read it on GitHub

Everyone is vibecoding. Nobody is verifying. Umbra scores it.

Umbra is a deterministic Trust Score (0–100) for AI-generated code: the vibe coding security scanner that verifies what your agent shipped, not what it claimed. One command, fully local, evidence for every finding.

→ umbra-badge.umbrabadge.workers.dev — the landing, live badges and report pages

Quickstart · Demo · The Audit · How it works · The Four Axes · FAQ · Roadmap · Contributing · Website

<!-- DEMO GIF: recorded from demo/demo.tape via charmbracelet/vhs. Specs:

npx @elberacasa/umbra . → verdict streams in → hold 3s on the final score

  • Terminal recording, 1200x600, dark theme
  • < 25 seconds total runtime
  • Beats per docs/demo-script.md: fresh shell → cd into vibe-coded app →

Re-record whenever the verdict format changes; stale demo output is a credibility bug (see docs/demo-script.md). -->

  • Render: cd demo && vhs demo.tape

Why Umbra exists

Studies put exploitable vulnerabilities in 40 to 60 percent of AI-generated code, and coding agents routinely claim "all tests pass" when three do. The tooling for writing code with AI is a year ahead of the tooling for trusting it. Umbra closes that gap: SAST rebuilt for how software gets written now, plus sandboxed verification that catches what static rules cannot.

One command scans any repo an agent produced (Claude Code, Cursor, Copilot, Windsurf, Lovable) and returns a score with file:line evidence for every finding. With --deep it goes further: Umbra builds and boots the repo in a locked-down Docker sandbox, then replays the agent's own claims against reality. If the agent is lying about tests, the score is capped below passing, with receipts.

The audit: 61 vibe-coded repos, scanned

We ran Umbra over 61 public, actively-maintained AI-built repos and published everything. The Vibe-Coding Security Audit:

Shortened. The full README is on GitHub.

Nothing above is checked by us. What we check is on the safety report.

Umbra: common questions

Is Umbra MCP server safe?
Mostly: it is graded B (80/100). Read the Umbra safety report
How do I install Umbra?
It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.
Does Umbra need an API key?
Not as far as the registry entry and our scan can tell: no credentials are declared or required.
Is Umbra maintained?
The last commit was 91 days ago (2026-08-05). The latest release is v1.2.0.
What can I use instead of Umbra?
Servers from other publishers that do the same job: LLM Sandbox MCP server, Basic Memory MCP server and HOL Guard MCP server.
How does Umbra compare with other AI and agents MCP servers?
safety score 80/100 (higher than 60% of 4,460 AI and agents servers); tools 3 (the typical AI and agents server has 6); tools that change things 0% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.

Alternatives to Umbra

Same job from other publishers: the closest match first, then the best rated.

  • LLM Sandbox
    Securely run LLM-generated code in isolated containers across 7 languages and 3 container backends.
    A
  • Basic Memory
    Local-first knowledge management with bi-directional LLM sync via Markdown files.
    B
  • HOL Guard
    Local-first AI agent security evidence and approval workflows through HOL Guard's stdio MCP server.
    A
  • OpenWorkProof
    Verifiable agent evidence; opt-in Codex guard and reference SOC alert review.
    A
  • Openflowkit
    Architecture diagrams for your agent: draw on the live OpenFlowKit canvas or edit diagram files.
    A