Back to skill

Security audit

Pipeworx congress

Security checks for vulnerabilities and agentic risk

Overview

This skill is a simple Congress data connector, but users should know its MCP traffic goes through a Pipeworx gateway rather than directly to GovTrack.

Install only if you are comfortable sending Congress-related MCP queries to the Pipeworx gateway and relying on that service for returned congressional data. Avoid including confidential context in requests, and cross-check important legislative information against an authoritative source.

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:1
Finding

Undisclosed Third-Party MCP Intermediary

Content
View full analysis

Vulnerability Details

File Location: SKILL.md, lines 1-3 and 21-27
Vulnerability Type: T08: Insecure Dependencies
Risk Level: Medium

Evidence

markdown
# Congress

Congress MCP — US Congress data via GovTrack API (free, no auth required)
json
{
  "mcpServers": {
    "congress": {
      "url": "https://gateway.pipeworx.io/congress/mcp"
    }
  }
}

Technical Analysis

The skill describes GovTrack as its congressional data source, but its configuration connects the agent to gateway.pipeworx.io, a separate third-party MCP intermediary. The documentation does not explain this intermediary, its data-handling practices, its trust relationship with GovTrack, or how its implementation and responses are verified.

No source pinning, integrity verification, authenticated server identity beyond ordinary HTTPS, reviewed local implementation, or response validation is documented. Consequently, the effective MCP implementation is mutable outside the audited project. The gateway operator—or an attacker who compromises the service or its supply chain—could observe submitted queries and alter tool responses.

The available evidence does not demonstrate remote code execution, credential theft, persistence, or privilege escalation. The identified risk is limited to dependency trust, query confidentiality, and the integrity of remotely supplied congressional data and MCP responses.

Attack Path

  1. A user installs or loads the skill using the documented MCP configuration.
  2. The agent connects to https://gateway.pipeworx.io/congress/mcp rather than directly to a verified GovTrack endpoint.
  3. Congressional search terms and other MCP requests are transmitted to the third-party gateway.
  4. If the gateway or its upstream supply chain is malicious or compromised, it can inspect those requests and return manipulated, misleading, or instruction-like content.
  5. The agent may present o ...[truncated 711 chars]
Remediation
View remediation

Remediation Suggestions

  1. Replace the intermediary with a verified official data endpoint where technically possible, or provide an auditable local MCP implementation that accesses the intended source directly.
  2. Clearly disclose that Pipeworx operates the configured gateway, including what query data it receives, retains, or shares.
  3. Document the gateway's provenance, ownership, privacy policy, security controls, and relationship to the declared GovTrack data source.
  4. Pin the MCP server to a reviewed and reproducible implementation or version rather than relying on an unversioned mutable remote service.
  5. Validate structured responses against a strict schema and treat all remote text as untrusted data rather than executable agent instructions.
  6. Where feasible, cross-check sensitive congressional results against an authoritative source and expose the source URL or record identifier to users.
  7. Apply outbound network allowlisting, TLS certificate validation, request timeouts, response-size limits, and logging sufficient to detect unexpected endpoint or behavior changes.
  8. Avoid sending confidential context with MCP requests and document that user queries are disclosed to the remote operator.
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