GovernClaw Policy Enforcer

• Outcome: Review
• Latest audit: May 29, 2026 · 9:26 PM UTC
• Version: 1.0.0

Overview

This governance skill is plausible, but it forwards full HTTP request details to a policy service and lets that service rewrite the request before sending it.

Review before installing. Use this only with a GovernClaw endpoint you operate or strongly trust, preferably local or otherwise authenticated and protected. Treat HTTP bodies and headers, including Authorization, Cookie, API keys, and tokens, as visible to GovernClaw, and be aware that the policy service can modify the final request before it is sent.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (4)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 89% confidence
• Finding: The skill declares no explicit permissions while clearly depending on environment access and outbound network communication. This weakens transparency and reviewability, making it easier for operators to deploy the skill without understanding that it can read configuration from the environment and send data to another service.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 92% confidence
• Finding: The skill markets itself as governance middleware for HTTP, shell, file, and browser actions, but the documented behavior only covers HTTP and also indicates direct external calls to a GovernClaw service. This mismatch can create a false sense of protection, causing users to believe sensitive shell, file, or browser actions are policy-enforced when they may not be governed at all.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The documentation states that request metadata such as session, agent, channel, and node identifiers are forwarded to GovernClaw, but it does not clearly present this as an outbound data-sharing behavior requiring user awareness. In a governance skill, this context can be sensitive operational metadata and may expose internal identifiers or workflow structure to another service.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The middleware sends the full governance payload, including session/agent identifiers and the requested URL, method, body, and headers, to the GovernClaw service before executing the action. In this skill context, that means potentially sensitive request contents and authorization headers are disclosed to another service by design, and there is no user-facing notice, minimization, or redaction to reduce privacy and secret-leak risk.

VirusTotal

66/66 vendors flagged this skill as clean.