ClawHub
Back to skill

Security audit

Social Content

Security checks for vulnerabilities and agentic risk

Overview

This skill provides social media content guidance and uses limited, purpose-related project context without signs of hidden execution or persistence.

Before installing, be aware that the skill may use local product marketing or brand voice files to tailor posts. Keep sensitive customer, financial, or confidential strategy details out of those files if you do not want them used in generated social content.

A.I.G

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Vague Triggers

Medium
Confidence
95% confidence
Finding
The skill description includes very broad trigger phrases such as 'social media,' 'engagement,' 'what should I post,' and 'grow my following,' which can match many ordinary requests and cause the skill to activate outside its intended scope. Over-broad activation is dangerous because it can route unrelated user tasks into a prompt that instructs the agent to read project files and apply domain-specific behavior, increasing the chance of unnecessary context access and unintended data exposure.

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The skill explicitly instructs the agent to read local files like `.agents/product-marketing-context.md` and `.agents/brand-voice-guide.md` before asking questions, but provides no user notice, consent step, or sensitivity check. This is risky because those files may contain proprietary strategy, brand rules, customer information, or other sensitive project data, and the skill normalizes automatic access based solely on invocation rather than user approval or least-privilege behavior.

Vague Triggers

Medium
Confidence
95% confidence
Finding
The eval explicitly rewards triggering on vague casual phrasing for a broad social-media skill without adding scope checks or disambiguation requirements. This can cause the router or agent to over-apply the skill to loosely related requests, increasing the chance of misrouting, bypassing better-specialized skills, and producing lower-quality or policy-inappropriate outputs.

Static analysis

No suspicious patterns detected.