ClawHub
Back to skill

Security audit

simulation-doc-writer

Security checks across malware telemetry and agentic risk

Overview

This is mostly a documentation-writing skill, but it asks the agent to create persistent diary files and propose changes to its own instructions after use.

Review before installing. The skill has clean VirusTotal telemetry and no executable code, but remove or ignore the Self-Evolution Mechanism unless you explicitly want diary files and PRs that propose changes to the skill itself. Also confirm whether you want .txt or .md output because the artifacts disagree.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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)

Description-Behavior Mismatch

Medium
Confidence
97% confidence
Finding
The self-evolution section instructs the skill to write diary files and alter its own operating rules after execution, which exceeds the declared purpose of producing simulation documentation. This creates unauthorized persistence and self-modification behavior, increasing the chance of unintended repository changes, instruction drift, and abuse through adversarial task inputs that trigger writes unrelated to the user’s documentation request.

Context-Inappropriate Capability

High
Confidence
99% confidence
Finding
The instruction to submit a PR modifying SKILL.md gives a documentation skill an unjustified capability to change its own definition and potentially expand its authority over time. In context, this is more dangerous because the skill handles untrusted user-provided content and could be induced to persist malicious or degraded rules into the codebase under the guise of self-improvement.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.