ClawHub
Back to skill

Security audit

twitter-viral-optimizer

Security checks across malware telemetry and agentic risk

Overview

The skill is a social-media writing assistant, but it asks for command execution without clear need or enforceable limits.

Review this skill carefully before installing. It does not show malicious exfiltration or destructive behavior, but you should only use it in an environment where command execution is disabled or tightly controlled, and provide platform API keys only through environment variables when a specific API-backed feature is actually needed.

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
94% confidence
Finding
The skill declares `exec` capability even though the document describes a largely markdown/LLM-driven social media optimization workflow and does not define concrete commands, inputs, or execution boundaries. This creates an unnecessary arbitrary command-execution surface: an agent may invoke shell commands based on user-controlled content or vague workflow steps, which can lead to file access, network pivoting, or environment secret exposure.

Intent-Code Divergence

Medium
Confidence
87% confidence
Finding
The skill states that no extra API key is required for core use, but nearby and later sections describe optional platform API keys and operational flows that expect authenticated API behavior. This inconsistency can mislead users and agents about what credentials may be needed, increasing the chance that secrets are requested ad hoc, mishandled, or supplied in unsafe ways during troubleshooting.

Intent-Code Divergence

Medium
Confidence
92% confidence
Finding
The security section claims only whitelist commands will be executed, but the skill provides no actual whitelist, wrapper, or enforcement mechanism. This kind of unenforced safety claim can cause operators to overtrust the skill while leaving the `exec` tool effectively unconstrained, enabling dangerous command execution if an agent or prompt interprets user input as shell operations.

Vague Triggers

Medium
Confidence
80% confidence
Finding
The description is broad and does not specify clear trigger conditions, scope limits, or disallowed operations beyond high-level marketing language. In agent environments, ambiguous invocation boundaries can cause over-application of the skill, including unnecessary use of privileged tools such as `exec` or handling of user data outside the intended content-optimization context.

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.