ClawHub
Back to skill

Security audit

ASCII Vision

Security checks for vulnerabilities and agentic risk

Overview

This skill is a local ASCII image-inspection helper with no evidence of hidden access, persistence, network use, or destructive behavior.

Install this if you are comfortable letting the agent run local ffmpeg and Python commands against images you choose. Treat the output as rough structural diagnostics, not full vision analysis or reliable object/text recognition.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

Tp4

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding
The core declared purpose of ASCII conversion without vision APIs is accurate, and the code does implement ASCII rendering plus basic brightness stats and a simple edge metric. However, the description overstates the implemented analysis capabilities: there is no color sampling at all because the script expects grayscale rawvideo input, and there is no explicit texture analysis beyond a simplistic adjacent-value threshold count. This is a description-to-behavior mismatch due to materially overstated capabilities, though there are no suspicious undeclared resource accesses or unrelated triggers.

Static analysis

No suspicious patterns detected.