T05 · Unauthorized Access and Privilege Escalation
Warning
- Location
SKILL.md:27- Finding
Excessive Filesystem and Command-Execution Permissions
- Content
View full analysis
- Remediation
View remediation
Security audit
模拟任意角色与经验级别的面试
Security checks for vulnerabilities and agentic risk
Overview
This interview simulator mostly describes mock interviews, but it asks for broad file and command powers that are not clearly needed or scoped.
Review this skill before installing. Its mock-interview examples are ordinary, but the declared read/write/exec tools could let an agent read or modify local files or run commands if the host grants those permissions. Use it only in a sandbox or with those tools disabled unless the publisher narrows the behavior to explicit user-selected interview materials and transcript export.
Vulnerability Patterns
- Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
- 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
Findings (1)
Vulnerability Patterns
- Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
- MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
- Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
- Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)
Description-Behavior Mismatch
High
- Category
- Not specified by scanner
- Confidence
- 97% confidence
- Finding
The skill is presented as a mock interview simulator, but the documentation and declared tools expand it into broad automation capabilities including file access, API use, and command execution. This creates a strong scope mismatch that can mislead users and downstream agents into granting or invoking powerful actions unrelated to the stated purpose, increasing the chance of unauthorized or unsafe operations.
- Content
Context-Inappropriate Capability
High
- Category
- Not specified by scanner
- Confidence
- 98% confidence
- Finding
Advertising system command execution is unjustified for a mock interview simulator and materially increases risk because an agent may invoke shell commands under the assumption they are part of normal skill behavior. In this context, exec capability is far more dangerous because the documented business purpose does not require it, so any use would be difficult for users to anticipate or validate.
- Content
Description-Behavior Mismatch
Medium
- Category
- Not specified by scanner
- Confidence
- 82% confidence
- Finding
The skill advertises support for custom question-bank uploads and saving/reviewing interview records, but no concrete interface, storage model, or data-handling controls are defined. Unsupported claims around data upload and retention are dangerous because users may assume private content will be stored or processed safely when the actual behavior and safeguards are unspecified.
- Content
Missing User Warnings
Medium
- Category
- Not specified by scanner
- Confidence
- 96% confidence
- Finding
The skill advertises file writing, API calls, and command execution without clear user warnings, approval boundaries, or impact disclosure. This is dangerous because users may invoke a seemingly harmless interview skill without understanding it could alter files, run commands, or transmit data, undermining informed consent and increasing the blast radius of misuse.
- Content
Context-Inappropriate Capability
Medium
- Category
- Not specified by scanner
- Confidence
- 95% confidence
- Finding
Broad file read/write capability is not justified by a mock interview simulator and can expose or modify local data unrelated to the interview task. In this context, the mismatch between purpose and capability makes the permission especially risky because users would not reasonably expect document access or file mutation from an interview practice tool.
- Content
Context-Inappropriate Capability
Medium
- Category
- Not specified by scanner
- Confidence
- 93% confidence
- Finding
Advertising external API integration is unjustified for a mock interview simulator and can lead to unnecessary data exfiltration or unintended third-party processing of user prompts and materials. Because the skill's core function can be performed locally by the model, undisclosed network access is a meaningful trust and privacy risk.
- Content
Natural-Language Policy Violations
Low
- Category
- Not specified by scanner
- Confidence
- 90% confidence
- Finding
The FAQ says the simulator 'mainly supports English,' which imposes a language preference in natural-language documentation. Because no user opt-in or clear justification for this locale limitation is provided, this conflicts with the language/locale policy criteria.
- Content
Static analysis
No suspicious patterns detected.