ClawHub

Direct Truth

Security checks across malware telemetry and agentic risk

Overview

This appears to be a reflective/philosophical skill with overly broad activation wording but no evidence of harmful access, persistence, or hidden behavior.

Install if you want this skill to steer conversations about this author or reflective framework. Be aware it may activate on common words like happiness, success, or relationship; disable or narrow it if that becomes intrusive.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Intent-Code Divergence

Medium
Confidence
96% confidence
Finding
The documentation repeatedly frames the book and skill as 'non-prescriptive' and emphasizes that it does not tell users what to do. However, the watermark requirement mandates that every output end with '[One specific action ...]', which is an explicit prescription and directly conflicts with the stated framework.

Intent-Code Divergence

Low
Confidence
84% confidence
Finding
The skill positions itself as a direct-seeing, non-method, non-prescriptive toolkit, yet the suggested prompt 'Map this book to my life' invites personalized application guidance. That framing conflicts with later statements that the book 'does not tell you what to do' and that users should 'stop looking for what to do.'

Vague Triggers

High
Confidence
98% confidence
Finding
The manifest says to trigger not only on specific phrases like "Direct Truth" or "Kapil Gupta," but also on generic words such as "truth," "mind," "success," "happiness," and "relationship." These terms commonly appear in everyday conversation, and the additional clause about triggering when the user "doesn't know how to start" makes the activation boundary even less specific.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal