ClawHub
Back to skill

Security audit

Rule Extractor

Security checks across malware telemetry and agentic risk

Overview

This skill is a small, disclosed rule-extraction helper, but users should review any generated prompt rules before reusing them.

Install only if you want a helper that turns traces and .learnings markdown titles into reusable agent rules. Review generated rules before adding them to a system prompt, and avoid running it over directories that may contain untrusted, private, or attacker-controlled notes.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • 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
88% confidence
Finding
The skill demonstrates file-read capability by extracting from the .learnings/ directory, but the metadata does not declare permissions or scope this access. In a skill that feeds extracted content into future system prompts, undeclared file access increases the risk of silently ingesting sensitive or attacker-planted data into higher-trust prompt context.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The description says rules are extracted and injected into later system prompts, but it does not warn users that untrusted trace or learning data can become privileged instructions in future sessions. This creates a prompt-persistence/prompt-injection pathway where malicious or erroneous content can be elevated into system-level behavior across conversations.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.