ClawHub
Back to skill

Security audit

Karpathy Compile

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it says: it turns local wiki notes into local knowledge-point files using a configured LLM, with a normal data-sharing caveat for the LLM endpoint.

Use this skill only with wiki notes you are comfortable sending to the configured LLM service. Keep the endpoint local or otherwise trusted for sensitive notes, review generated knowledge points before relying on them, and avoid running the bundled end-to-end tests unless you have also reviewed the sibling pipeline skills they load.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (1)

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The code sends aggregated wiki content to an LLM endpoint (`http://localhost:11434/v1`) without any consent, disclosure, classification, or filtering step. Even though the default endpoint is local, this still creates a data-exposure path because wiki entries may contain sensitive internal content, and the endpoint is configurable, so data could be redirected to a remote service or logged by another local process.

Static analysis

No suspicious patterns detected.