ClawHub
Back to skill

Security audit

Context Driven Development

Security checks across malware telemetry and agentic risk

Overview

This skill appears development-oriented, but it asks for shell execution authority without clear limits or user-control guidance.

Review this before installing if you will use it on sensitive projects. It does not show malicious behavior, but because it grants shell execution in a broadly scoped development workflow, users should require explicit approval before commands run and avoid using it on repositories where unintended tests, deployment commands, or credential exposure would be costly.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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
Findings (4)

Context-Inappropriate Capability

Medium
Confidence
91% confidence
Finding
The skill declares the exec tool even though its stated purpose is context/document management and code assistance, without clearly constraining what commands may be run. This expands the attack surface by enabling shell execution under a vague scope, which could lead to unintended command execution or abuse when the agent interprets user/project content as actionable instructions.

Intent-Code Divergence

Medium
Confidence
87% confidence
Finding
The document primarily frames the skill as context/document management, but later treats command execution failure as a normal operating condition. That mismatch can mislead users and agents about the operational behavior of the skill, increasing the chance that shell execution occurs without informed consent or adequate safeguards.

Vague Triggers

Medium
Confidence
80% confidence
Finding
The activation guidance is broad and ambiguous, covering code generation, debugging, testing, and deployment without precise triggers or boundaries. In combination with read and exec capabilities, this ambiguity can cause over-invocation in sensitive contexts and increase the likelihood of risky actions being taken based on loosely matched requests.

Missing User Warnings

Medium
Confidence
93% confidence
Finding
The skill exposes command execution capability but does not warn users in the Markdown description that shell commands may be run. This is dangerous because users may supply untrusted project context believing the skill is documentation-oriented, while the agent may legitimately execute commands as part of the skill.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.