Back to skill

Security audit

Pipeworx agify

Security checks for vulnerabilities and agentic risk

Overview

The skill appears purpose-aligned for querying an external age-prediction service, but its MCP setup uses an unpinned package that can execute changing third-party code locally.

Review before installing. Use only for non-sensitive first-name age-prediction queries, assume submitted names are sent to Pipeworx and may reach Agify, and prefer a pinned reviewed MCP bridge version instead of `mcp-remote@latest`.

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

Warning
Location
SKILL.md:46
Finding

Unpinned Third-Party Package Is Downloaded and Executed Automatically

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 46–49
Vulnerability Type: Unpinned npm dependency and automatic remote package execution
Risk Level: Medium

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

Technical Analysis

The MCP configuration invokes npx with the -y option and references mcp-remote@latest. The -y option suppresses installation confirmation, while the @latest tag dynamically resolves to whichever package release is current at execution time.

Consequently, the code executed by this configuration is not fixed to the version reviewed during the audit. A future malicious or compromised release of mcp-remote, compromise of its npm publishing account, or compromise elsewhere in the package distribution chain could cause arbitrary third-party code to run locally. The effective behavior can therefore change without any modification to SKILL.md.

Attack Path

  1. An attacker compromises the npm package, its publisher account, or a dependency included in a future mcp-remote release.
  2. The malicious release becomes the version resolved by the latest distribution tag.
  3. A user or MCP client loads the documented configuration.
  4. The client runs npx -y mcp-remote@latest.
  5. npx downloads the newly resolved package without requesting installation approval.
  6. Package or lifecycle code executes with the privileges and environment of the user running the MCP client.

Impact Assessment

Successful exploitation could permit arbitrary code execution under the MCP client's operating-system user account. Depending on that user's permissions and environment, malicious package code could read or modify accessible files, inspect environment variables and credentials, alter MCP communications, initiate network requests, or execute addi ...[truncated 303 chars]

Remediation
View remediation

Remediation Suggestions

  • Replace mcp-remote@latest with an explicitly pinned and reviewed version, such as mcp-remote@x.y.z.
  • Remove -y where interactive installation approval is practical.
  • Install dependencies through a controlled build or deployment process rather than downloading them dynamically whenever the MCP client starts.
  • Use a lockfile and npm integrity hashes to ensure the resolved package and transitive dependencies match reviewed artifacts.
  • Configure an approved package registry and apply package provenance or signature verification where supported.
  • Scan the pinned package and its transitive dependencies before deployment.
  • Establish a controlled update process in which new versions are reviewed and tested before the pinned version is changed.
  • Run the MCP bridge with least privilege and restrict its access to sensitive files, credentials, and unnecessary network destinations.
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 9)May include surrounding context.

md
openclaw:
    requires:
      bins:
        - curl
    emoji: "⚡"
    homepage: https://pipeworx.io/packs/agify
---

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill instructs the agent to send user-supplied data to an external HTTP endpoint but does not warn that names and query contents leave the local environment. This creates a privacy and consent issue because users may provide personal data without realizing it will be transmitted to a third-party gateway and underlying service.

Content

No source excerpt is available for this finding.

Vague Triggers

Low
Category
Not specified by scanner
Confidence
90% confidence
Finding

The phrase "Use this skill when the user asks about agify data or needs to query agify information" is circular and does not define concrete trigger phrases, boundaries, or exclusion conditions. This can lead to ambiguous activation because it does not clearly distinguish when the skill should be invoked versus when a general information request should not trigger it.

Content

No source excerpt is available for this finding.