ClawHub
Back to skill

Security audit

Local Coding

Security checks for vulnerabilities and agentic risk

Overview

This is a documentation-only local AI coding skill with disclosed local routing and no evidence of hidden, destructive, or exfiltrating behavior.

Reasonable to install if you want a local AI router for coding. Verify the `ollama-herd` package and publisher before installing, use it only on trusted local networks, and avoid sending secrets in prompts because local traces or logs may remain under `~/.fleet-manager/`.

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)

Description-Behavior Mismatch

Medium
Confidence
94% confidence
Finding
The skill is marketed as a local coding assistant, but the documented capabilities also include general-purpose chat, image generation, and speech-to-text. This scope expansion can mislead users and higher-level agents into granting broader access than expected, increasing the chance of unsafe or unintended use of the local model fleet.

Static analysis

No suspicious patterns detected.