ClawHub
Back to skill

Security audit

Strategic Commitment

Security checks across malware telemetry and agentic risk

Overview

This is a transparent strategy-coaching skill with no executable code or hidden data access, though users should apply it only in lawful and non-coercive contexts.

Install only if you want a game-theory negotiation and strategy framework. Do not use it to coerce, intimidate, exploit, evade law, or make irreversible commitments without appropriate legal, ethical, and downside review.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (2)

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill teaches users how to make threats or promises more credible by changing payoff structures and explicitly discusses deterrence, lock-in, and making retreat more expensive, but it provides no safety framing about coercion, abuse, legal/ethical limits, or irreversible harm. In an agent setting, this can normalize manipulative escalation and enable harmful real-world conduct, especially when applied to interpersonal, commercial, or adversarial situations without guardrails.

Missing User Warnings

Medium
Confidence
94% confidence
Finding
This section operationalizes commitment devices such as sunk cost, third-party constraints, and organizational lock-in, and instructs the user to raise the cost of retreat until retreat is no longer dominant. Encouraging irreversible actions without an explicit warning or constraint is dangerous because it can be directly repurposed for coercive bargaining, anticompetitive conduct, self-harmful overcommitment, or escalation strategies that are hard to unwind once executed.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.