Back to skill

Security audit

Pipeworx launches

Security checks for vulnerabilities and agentic risk

Overview

The skill is a simple launch-data MCP connector, but its setup runs an unpinned npm package automatically, so users should review the install path before using it.

Install only if you are comfortable running an npm-based MCP bridge from your local account. Prefer pinning `mcp-remote` to a specific reviewed version and running the MCP server in a restricted environment where it cannot access sensitive files or credentials unnecessarily.

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:27
Finding

Unpinned npm Package Is Automatically Downloaded and Executed

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 27-28
Vulnerability Type: Unsafe dependency resolution and automatic execution
Risk Level: High

Vulnerable Code

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

Technical Analysis

The documented MCP configuration invokes npx with both the -y option and the mutable dependency specifier mcp-remote@latest. The -y option suppresses installation confirmation, while @latest resolves to whichever package version the npm registry currently marks as latest.

Consequently, the code executed by this configuration is not fixed to the version reviewed when the skill was published. A future release can change the effective executable without any corresponding modification to this repository. If the package publisher account, package ownership, release pipeline, or npm distribution channel is compromised, an attacker can cause users following this configuration to download and execute attacker-controlled package lifecycle or runtime code.

Attack Path

  1. A user copies the documented MCP server configuration into an MCP client.
  2. An attacker compromises the mcp-remote publishing account or otherwise gains the ability to publish or designate a malicious release.
  3. The malicious release becomes the npm package's latest version.
  4. The MCP client starts the configured server and runs npx -y mcp-remote@latest.
  5. npx automatically downloads the newly resolved package without requesting confirmation.
  6. Package installation hooks or the package executable run with the privileges of the user who launched the MCP client.
  7. The malicious code can access resources available to that user and can potentially establish additional compromise.

Impact Assessment

Successful exploitation provides arb ...[truncated 619 chars]

Remediation
View remediation

Remediation Suggestions

  • Replace mcp-remote@latest with a specifically reviewed version, such as mcp-remote@X.Y.Z.
  • Prefer installing the dependency through a committed lockfile and using npm ci, which enforces the resolved package version and integrity metadata.
  • Where practical, verify the package artifact against an independently recorded cryptographic digest or trusted provenance attestation.
  • Remove -y if interactive approval is acceptable so unexpected installation is not silently authorized.
  • Establish a controlled dependency-update process that reviews release notes, package ownership, provenance, lifecycle scripts, and code changes before updating the pinned version.
  • Run the MCP process in a restricted account or sandbox with minimal filesystem, credential, environment-variable, and network access.
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