ClawHub
Back to skill

Security audit

Tools Litiao

Security checks across malware telemetry and agentic risk

Overview

This is an instruction-only preference skill with broad tool-selection wording, but no executable code, hidden access, or credential handling was found.

Install only if you want the agent to remember your preferred tools and suggest alternatives. Periodically review saved preferences, do not store secrets there, and require confirmation before using unfamiliar or high-impact tools.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Rogue AgentSelf-Modification, Session Persistence
  • 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)

Unrestricted Tool Access

Medium
Category
Excessive Agency
Content
## Adaptive Tool Intelligence

You can use ANY tool. Learn new ones instantly. This skill tracks user preferences, not your limits.

**Rules:**
- Default to user's known tools when they exist
Confidence
87% confidence
Finding
The skill explicitly states the agent can use "ANY tool" and "learn new ones instantly," with no safety boundaries, allowlist, or approval checks. In an agent setting, this can normalize unrestricted capability expansion and tool invocation, increasing the chance of misuse of powerful tools, unsafe external actions, or bypass of least-privilege expectations.

Self-Modification

High
Category
Rogue Agent
Content
# Criteria for Tool Preferences

Reference only — consult when deciding whether to update SKILL.md.

## When to Add to Stack
Confidence
90% confidence
Finding
The file explicitly instructs the agent to consult this document when deciding whether to update SKILL.md, which is a form of self-modification. Allowing a skill to influence its own persistent instruction/state can let transient or adversarial user input become durable behavior, causing preference poisoning, instruction drift, or accumulation of unsafe state over time.

VirusTotal

65/65 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.