Back to skill

Security audit

PPT Pandora

Security checks for vulnerabilities and agentic risk

Overview

This skill is a PPT production guide with proportionate local editing, image review, and sharing examples, with no evidence of hidden or malicious behavior.

Install only if you want a Chinese-oriented PPT production workflow. Keep the agent limited to the intended presentation folder, review before it deletes old PPT versions or media, and confirm the Feishu recipient before sending files.

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
  • 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 (4)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The example hard-codes the run language as zh-CN, which instructs generated PPT content to use a specific locale by default. Under the policy, forcing a language or locale without offering user choice is a natural-language policy violation unless it is explicitly justified as region-specific.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
82% confidence
Finding

At L612-L633 the document states that image functionality is not recommended for automation and that image insertion should remain manual. However, L637-L695 then documents automated image-quality detection and identifies unreferenced media as 'safe to delete', which is a materially different automated capability over image assets. This is an intent-level contradiction within the skill documentation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The example hard-codes lang: 'zh-CN' when creating PPT text runs, which forces a specific language/locale in output behavior. The file does not indicate that users can choose another locale or that this restriction is limited to a documented region-specific use case.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

L0613-L0627 明确写明“图片功能暂不支持自动化部署”,并建议“图片插入:手动添加,AI只负责提供图片建议和位置”。但前文多处代码示例实际演示了自动插入/替换图片能力,例如 L0067-L0085 的 add_picture、L0173-L0189 的图片替换工作流,形成了文档意图与实现指导之间的直接冲突。

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.