Back to skill

Security audit

Pipeworx npm

Security checks for vulnerabilities and agentic risk

Overview

This skill appears purpose-aligned, but its setup runs an unpinned npm package automatically, so users should review the install risk before enabling it.

Install only if you are comfortable running the Pipeworx remote MCP bridge locally. Prefer pinning `mcp-remote` to a reviewed exact version and running the MCP process with least-privilege filesystem, environment, and network access.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Findings (1)

T08 · Insecure Dependencies

Error
Location
SKILL.md:23
Finding

Unpinned npm Package Is Automatically Downloaded and Executed

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 23-29
Vulnerability Type: T08: Insecure Dependencies
Risk Level: High

The Skill configuration directs npx to download and execute the mutable latest release of mcp-remote without version or integrity pinning:

json
{
  "mcpServers": {
    "pipeworx-npm": {
      "command": "npx",
      "args": ["-y", "mcp-remote@latest", "https://gateway.pipeworx.io/npm/mcp"]
    }
  }
}

Technical Analysis

The @latest npm tag is mutable and may resolve to different package contents each time the MCP server starts. The -y option suppresses the interactive installation confirmation, causing the resolved package to be downloaded and executed automatically.

This creates a supply-chain risk because the effective executable code is not fixed to the version reviewed during the audit. If the upstream package, maintainer account, publishing credentials, npm distribution channel, or mutable tag is compromised, a malicious release can be delivered to future users without any change to this Skill file.

Attack Path

  1. An attacker compromises the mcp-remote package publication process or otherwise causes the npm latest tag to reference a malicious release.
  2. A user configures or launches the Skill using the documented MCP configuration.
  3. npx -y resolves mcp-remote@latest, downloads the attacker-controlled release, and suppresses installation confirmation.
  4. The downloaded package executes locally under the account running the Agent.
  5. Malicious package lifecycle or runtime code performs actions available to that account.

Impact Assessment

Successful exploitation permits arbitrary code execution with the privileges of the user running the Agent or MCP server. Depending on that user's permissions, the malicious dependency could access readable local files and environment variables, modify user-owned data or configuration, make ...[truncated 279 chars]

Remediation
View remediation

Remediation Suggestions

  • Replace mcp-remote@latest with an explicitly reviewed, exact package version, such as mcp-remote@x.y.z.
  • Avoid automatic execution of dynamically resolved packages. Install the dependency through a controlled build or deployment process and invoke the installed binary.
  • Commit and enforce a lockfile where applicable, and verify package integrity using npm integrity metadata or an equivalent cryptographic verification mechanism.
  • Review package provenance, maintainers, release history, and published artifacts before approving version updates.
  • Route dependency updates through automated security scanning and human review rather than tracking a mutable distribution tag.
  • Run the MCP process in a least-privilege sandbox with restricted filesystem access, environment variables, process execution, and outbound network access to reduce the consequences of a compromised dependency.
Vulnerability Patterns
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep