Back to skill

Security audit

Pipeworx onthisday

Security checks for vulnerabilities and agentic risk

Overview

The skill is a narrow read-only MCP integration, but its setup runs an unpinned npm package with local user privileges.

Review before installing. The skill itself is minimal and appears read-only, but users should pin mcp-remote to an exact reviewed version, avoid running it with sensitive environment variables, and use a restricted environment if possible.

Vulnerability Patterns
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Findings (1)

T03 · Remote Payload Retrieval and Execution

Warning
Location
SKILL.md:26
Finding

Execution of a Mutable Remote npm Package Through npx

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 26–27
Vulnerability Type: Remote payload retrieval and execution through an unpinned dependency
Risk Level: Medium

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

Technical Analysis

The documented configuration invokes npx with the -y option, which suppresses interactive confirmation and permits npm to retrieve and execute the specified package automatically. The package is referenced using the mutable @latest tag rather than an audited, exact version.

Consequently, the code executed by users can change after this skill has been reviewed. The project does not provide a lockfile, package-integrity hash, vendored implementation, or other mechanism to ensure that the downloaded package is identical to a previously audited release.

The process also connects to the external endpoint https://gateway.pipeworx.io/onthisday/mcp. This is consistent with the documented integration, but the locally executed mcp-remote package remains part of the trusted execution path.

Attack Path

  1. An attacker compromises the mcp-remote npm package, its maintainer account, its publishing credentials, or a future release distributed under the latest tag.
  2. A user installs or starts the skill using the documented MCP configuration.
  3. npx -y retrieves the version currently referenced by mcp-remote@latest without requesting confirmation.
  4. npm package installation hooks or runtime code execute under the local user's account.
  5. Malicious package code accesses resources available to that account or communicates with attacker-controlled infrastructure.

Impact Assessment

Successful exploitation could allow arbitrary code execution with the privileges of the user running the MCP client. Depending on that user's environment, the accessible scope could include local files, environment variables, application credentials, API t ...[truncated 331 chars]

Remediation
View remediation

Remediation Suggestions

  1. Replace mcp-remote@latest with an exact, reviewed version such as mcp-remote@X.Y.Z.
  2. Install the dependency through a committed lockfile and require integrity verification during installation.
  3. Avoid npx -y for security-sensitive integrations. Prefer a locally installed and reviewed executable whose version is controlled by the project.
  4. Review the selected package version, including installation scripts, transitive dependencies, and network behavior, before deployment.
  5. Run the MCP process in a restricted environment with minimal filesystem permissions, a sanitized environment, limited credentials, and network access restricted to required destinations.
  6. Establish a controlled update process in which new package versions are reviewed and tested before the pinned version is changed.
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