ClawHub
Back to skill

Security audit

Skill Evolver

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed skill-management workflow, but users should approve any installs, audits, or new skill creation deliberately.

Install this only if you want an agent to help manage other skills. Review any registry skill before installing it, be cautious with global npm/pnpm or npx commands, verify selected skill slugs and sources, inspect generated fused skills before reuse, and do not run any generated removal command unless the path is exactly the intended skill directory.

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 (2)

Lp3

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding
The skill instructs the agent to read and write files and execute shell commands (`mkdir -p`, generating reports in `output/...`) without declaring permissions. This creates a transparency and governance gap: operators may believe the skill is passive/orchestration-only while it can mutate the filesystem and invoke the shell, increasing the risk of unsafe execution or misuse in environments that rely on declared permissions for policy enforcement.

Tp4

High
Category
MCP Tool Poisoning
Confidence
89% confidence
Finding
The stated purpose is lifecycle/orchestration of skills, but the workflow also performs installation, verification, security auditing, filesystem scanning, and report generation. This mismatch is dangerous because users and automated routing may invoke the skill under a lower-risk mental model, while the actual behavior expands into security-sensitive inspection and system-modifying operations, which can lead to over-trust and unintended access or execution paths.

VirusTotal

60/60 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.