ClawHub
Back to skill

Security audit

Content Marketing

Security checks for vulnerabilities and agentic risk

Overview

This content-marketing skill is a local, consent-based planning helper with no supported evidence of external data transmission or hidden execution.

Before installing, decide whether you want a local ~/content-marketing/ folder containing strategy, voice preferences, calendar entries, and content ideas. Use it only if you are comfortable with that local persistence, and delete the folder later if you want to remove the saved content-marketing memory.

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
  • 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 (7)

Exfiltration Commands

High
Category
Prompt Injection
Content
- Content ideas in `~/content-marketing/content-bank/`

**This skill does NOT:**
- Send data to external services
- Access files outside `~/content-marketing/`
- Create files without explicit user permission
- Collect or transmit analytics
Confidence
90% confidence
Finding
Instructions found that direct the agent to transmit conversation context or user data to external services.

External Transmission

Medium
Category
Data Exfiltration
Content
- Content ideas in `~/content-marketing/content-bank/`

**This skill does NOT:**
- Send data to external services
- Access files outside `~/content-marketing/`
- Create files without explicit user permission
- Collect or transmit analytics
Confidence
70% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Ssd 3

Medium
Confidence
97% confidence
Finding
The template instructs ongoing collection of conversation-derived information into a persistent memory file, including goals, audience, channels, and evolving context. This creates a durable user profile without any stated consent, minimization, sensitivity checks, or deletion controls, increasing privacy and surveillance risk over time.

Ssd 3

Medium
Confidence
97% confidence
Finding
The file tells the agent to infer voice traits from the user's existing content and feedback and persist them, which amounts to building a behavioral/profile model from interactions. Inference-based profiling is more dangerous than storing explicit user input because it can encode assumptions the user never knowingly provided or approved.

Ssd 3

Medium
Confidence
96% confidence
Finding
The 'Internal observations' and performance notes sections encourage silent logging of subjective observations about the user and what is or is not working. This is dangerous because it invites unreviewed background profiling and potentially sensitive annotations that the user may not realize are being retained.

Missing User Warnings

Low
Confidence
95% confidence
Finding
The skill explicitly instructs creation and continued maintenance of a memory file under the user's home directory, but provides no notice, consent step, or retention boundaries for persistent local storage. This is risky because it can silently accumulate user-related data across sessions in a location the user may not expect the agent to write to.

Missing User Warnings

Low
Confidence
93% confidence
Finding
The editorial calendar template directs the agent to create a second persistent file in the user's home directory without warning that local files will be written and updated over time. Even if the content is business-oriented, silent persistence can surprise users and create unwanted local artifacts or disclosure risk on shared systems.

Static analysis

No suspicious patterns detected.