Security audit

AI KEY·做成一个技能

Security checks for vulnerabilities and agentic risk

Overview

The skill saves proven workflows as local skills with clear publishing limits, but relies on companion files absent from this package.

Before use, inspect the companion configuration, shared rules, and listing script, which were not bundled and remain unverified. Expect local skill changes, evaluation files, and workflow memory; review saved material for private details. Publishing requires your explicit request, and generated descriptions will include Chinese trigger phrases.

A.I.G

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

SkillSpector

Vulnerability Patterns

Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
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 (1)

Generated skill descriptions must include Chinese trigger phrases without user opt-in.

Low

Category: Output Handling
Confidence: 96% confidence
Finding: The self-check requires the generated description to contain Chinese trigger phrases (「含中文触发词」), regardless of the user's preferred language. No language choice or justification for this mandatory locale requirement is provided.
Content: No source excerpt is available for this finding.