ClawHub
Back to skill

Security audit

Gangtise Agent调用

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed Gangtise research API helper that needs Gangtise credentials, with some credential-handling cautions but no hidden or destructive behavior found.

Install only if you intend to use Gangtise credentials with this skill. Prefer environment-provided secrets over a plaintext scripts/.authorization file, do not commit credential files, restrict local file permissions, and avoid setting GANGTISE_OPENAI_ROOT to any untrusted endpoint because authorization headers are sent there.

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
92% confidence
Finding
The skill instructs users to provide long-lived access credentials via environment variables or a local file, but does not warn about secret handling, file permissions, accidental commits, or leakage through logs and outputs. This increases the chance that access keys are stored insecurely or exposed to other tools, users, or source control, leading to unauthorized API use.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The code reads long-term credentials from environment variables or a local file and sends them to a remote service to mint an access token, but there is no user-facing disclosure, timeout, or stronger handling around this sensitive flow. In an agent-skill context, hidden outbound use of stored secrets increases the risk of unintended credential exposure and makes trust decisions harder for users and operators.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.