Security audit

Exam Evaluator

Security checks for vulnerabilities and agentic risk

Overview

This appears to be a legitimate exam-analysis skill, with minor caveats around broad document triggers, sensitive local report data, and missing helper scripts.

Install only if you intend to analyze exam or question-bank documents. Check uploaded files for sensitive personal or institutional data, verify the referenced parsing scripts before running them, and delete temp/output reports when finished; for highly sensitive reports, consider avoiding external CDN-loaded chart scripts.

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
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 (2)

Vague Triggers

Medium

Confidence: 88% confidence
Finding: The trigger list is broad enough to match routine educational conversations and generic file mentions such as .xlsx/.docx/.pdf, which can cause the skill to activate outside the user's actual intent. In an agent system, over-broad activation can route unrelated documents into a workflow that parses local files, asks forced follow-up questions, and generates reports, increasing the chance of unintended data handling and confusing or unsafe tool use.

Natural-Language Policy Violations

Medium

Confidence: 80% confidence
Finding: The skill description is written to operate in Chinese without offering language negotiation or documenting that it is restricted to Chinese-language exam materials. This can cause misinterpretation of user intent, incorrect parsing/evaluation of non-Chinese content, and misleading outputs when the user interacts in another language.

Static analysis

No suspicious patterns detected.