ClawHub

felt cute uploaded skill

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed codebase research helper that reads repository content, runs basic git metadata commands, uses researcher subagents, and writes a local report.

Install only if you are comfortable with the agent reading the relevant codebase, spawning parallel researcher subagents, running basic git metadata commands, and writing research.md in the current directory. Check for an existing research.md first, and make sure your git remote URL does not contain credentials before sharing the report.

SkillSpector

By NVIDIA
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 (2)

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill explicitly instructs writing `research.md` in the current working directory and says to always produce it, but does not require checking whether the file already exists or obtaining user confirmation before overwriting. In an agent context, this can cause unintended modification or loss of user data, especially when the working directory already contains a file with that name or when the user expected read-only research behavior.

Missing User Warnings

Low
Confidence
84% confidence
Finding
The skill directs collection of repository metadata including `git remote get-url origin`, which may expose internal repository URLs, embedded credentials, private hostnames, or organization names. While this is not inherently malicious, automatically gathering and potentially placing this value into `research.md` increases the risk of leaking sensitive infrastructure details into outputs that may be shared more broadly.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal