ClawHub
Back to skill

Security audit

AI KEY·重讲一个概念

Security checks for vulnerabilities and agentic risk

Overview

This writing skill coherently supports drafting and editorial follow-up, with disclosed file updates and calendar reminders.

Choose the content vault carefully. Expect persistent editorial updates and 24-hour/7-day calendar reminders after publication; explicitly limit the run if you only want a draft. Review the separately supplied family documents and companion skills, whose behavior could not be assessed from this bundle.

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
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (1)

The workflow extends concept-script drafting into persistent post-publication scheduling and edits to a subsequent script.

Medium
Category
Excessive Agency
Confidence
87% confidence
Finding

The manifest describes explaining a concept and producing a script finalized through the host's spoken take. These instructions additionally require publication-log updates, T+24h/T+7d calendar creation, and changes to the next script's frontmatter—persistent operations beyond producing the current concept script.

Content

No source excerpt is available for this finding.