Back to skill

Security audit

bookreading

Security checks for vulnerabilities and agentic risk

Overview

This is a narrow Chinese book-learning prompt that asks the agent to search public book information and summarize it, with no hidden execution, credential use, persistence, or data access.

Install this if you want a Chinese-language assistant for studying books from public source material. Be aware it relies on search results, so obscure books may produce incomplete summaries, and users should avoid requesting long copyrighted excerpts from books.

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
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 (2)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest description is entirely prescriptive in Chinese, including requirements such as '语言简洁', and the skill is presented as the default behavior rather than an opt-in locale-specific tool. This creates a language policy concern because the skill appears to mandate a specific language without giving the user a choice or explaining that it is intended only for Chinese-language use.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
80% confidence
Finding

These instructions impose a specific language/style regime as a mandatory rule, but the document does not provide a user-selectable language option. While style guidance is acceptable, coupling it with a Chinese-only skill context reinforces an undocumented locale constraint.

Content

No source excerpt is available for this finding.