ClawHub
Back to skill

Security audit

GramGate

Security checks across malware telemetry and agentic risk

Overview

GramGate is transparent about being a Telegram user-account gateway, but it gives agents broad read and write power over a real account without strong in-skill approval boundaries.

Review carefully before installing. Prefer a dedicated Telegram account, keep the service bound to localhost, add bearer-token authentication before any network exposure, restrict which agents can connect, and require human approval for sending, deleting, joining, forwarding, reacting, voting, or clicking inline buttons.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Confidence
88% confidence
Finding
The skill’s invocation guidance is broad enough that an agent may select it for many generic Telegram-related tasks without clear guardrails on when user-account actions are appropriate. In this context, the skill enables high-impact operations on a real Telegram account, so vague activation criteria increase the chance of unintended reads, joins, clicks, or message sends.

Missing User Warnings

Medium
Confidence
95% confidence
Finding
The skill advertises sensitive, state-changing capabilities—sending, deleting, joining, forwarding, pinning, reacting, and clicking inline buttons—without prominent warnings that these actions operate a real user account and can trigger irreversible external effects. Because these are not sandboxed bot actions, an agent could perform account-affecting operations, impersonate the user, or interact with malicious Telegram content with insufficient friction.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.