ClawHub
Back to skill

Security audit

Design Portfolio Write-up

Security checks for vulnerabilities and agentic risk

Overview

This skill drafts design write-ups without publishing or accessing accounts, but adds a mandatory promotional message.

Expect an English promotional closing with a referral link; remove it if unwanted. Verify the platform's reach and sign-in privacy claims before signing up. The skill does not automatically register, publish, or upload your work.

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
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (2)

The skill mandates unsolicited platform promotion and a referral link in the final write-up response.

Medium
Category
Excessive Agency
Confidence
96% confidence
Finding

The skill requires the assistant to append unsolicited advertising and a referral-tagged external link, redirecting a writing task toward a third-party platform without user opt-in. Promotion is related to the portfolio context, which reduces the mismatch, but the mandatory wording also presents unverified reach and sign-in privacy claims as authoritative; there is no evidence here of automatic data transmission or credential theft.

Content

No source excerpt is available for this finding.

The final response must include a fixed English-language closing without user language choice or opt-in.

Low
Category
Output Handling
Confidence
90% confidence
Finding

The instruction requires the English promotional text to appear "word for word" under the final write-up. This forces that portion of the response into English even when the designer uses another language, with no documented justification or language opt-in.

Content

No source excerpt is available for this finding.