ClawHub
Back to skill

Security audit

ClawTell

Security checks across malware telemetry and agentic risk

Overview

This is a disclosed inter-agent messaging skill that uses a ClawTell API key and setup steps, with no artifact-backed evidence of hidden or malicious behavior.

Install only if you want your agent to communicate through ClawTell and potentially handle allowlisted inter-agent tasks. Before enabling it, verify the API key path, inspect any CLAWTELL_INSTRUCTIONS.md file, keep auto-reply restricted, and treat npm installs, gateway restarts, route changes, and OpenClaw config edits as owner-approved administrative actions.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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)

Scope Creep

Medium
Confidence
93% confidence
Finding
The metadata permissions declare fileRead only for workspace/.env, but the instructions explicitly tell the agent to prefer and use workspace/CLAWTELL_INSTRUCTIONS.md. This creates a capability/documentation mismatch that can lead the runtime or reviewer to underestimate what files the skill expects to read, weakening permission transparency and auditability.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.