T01 · Skill Instruction Hijacking
Warning
- Location
SKILL.md:20- Finding
Untrusted Job Description Can Inject Instructions into the Parsing Prompt
- Content
View full analysis
- Remediation
View remediation
Security audit
51mee Position Parse
Security checks for vulnerabilities and agentic risk
Overview
This is a simple job-description parsing skill with minor prompt and output-format issues, but no evidence of code execution, persistence, credential access, or data theft.
Installers should treat this as a low-risk parsing helper. For stronger reliability, run it in a low-privilege context and improve the prompt to treat the job description strictly as untrusted data, enforce the JSON schema, and resolve the conflicting output-format instructions.
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
Findings (1)
Vulnerability Patterns
- Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
- 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)
Vague Triggers
Low
- Category
- Not specified by scanner
- Confidence
- 86% confidence
- Finding
The trigger text says the skill applies when a user provides a job description for parsing or wants to analyze a JD's core requirements, but it does not define boundaries or negative examples for when similar requests should not invoke this skill. In a manifest/markdown context, this can lead to overlap with more general resume, hiring, or document-analysis requests.
- Content
Intent-Code Divergence
Low
- Category
- Not specified by scanner
- Confidence
- 93% confidence
- Finding
SKILL.md 在“输出格式”部分定义了一个完整的 Markdown 报告模板（L083-L112），表明技能应输出人类可读报告；但“注意事项”又明确要求“直接返回 JSON，不要额外解释”（L118），与前述输出格式相冲突。这属于文档对实现意图的主动矛盾，会让调用方无法确定技能究竟应返回结构化 JSON 还是 Markdown 报告。
- Content