ClawHub
Back to skill

Security audit

Agent Budget Governance

Security checks across malware telemetry and agentic risk

Overview

This skill appears local and non-exfiltrating, but it overstates its enforcement abilities and can persistently rewrite multi-agent budget governance state without adequate controls.

Review this before installing in a real multi-agent environment. Treat it as an advisory local budget-file auditor, not proven spawn-privilege enforcement. Back up BUDGET.json files, verify your schema and timezone assumptions, and do not schedule it automatically until you add a real dry run, rollback plan, and confirmed enforcement for privilege revocation.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
90% confidence
Finding
The skill advertises local file operations in metadata but does not declare permissions in a way that clearly exposes those capabilities to users or policy systems. Undeclared read/write access is dangerous because the skill is explicitly designed to modify per-agent budget files and governance logs, which can affect agent behavior and operational controls without clear consent boundaries.

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding
The documentation claims governance features such as mesh privilege revocation and trust-based demotion, but the described implementation appears limited to local budget-file mutation and status tracking. This mismatch is dangerous because operators may rely on the skill for enforcement that does not actually occur, creating a false sense of security while overspending or privilege abuse remains unmitigated.

Missing User Warnings

Medium
Confidence
86% confidence
Finding
The skill documentation describes modifying budget state and revoking agent privileges without an explicit warning that these are system-affecting operations. In a multi-agent environment, silently changing governance files or spawn permissions can disrupt workflows, lock out agents, or cause operators to run the skill without understanding the operational consequences.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.