Security audit
验证模式 (Verify Mode)
Security checks for vulnerabilities and agentic risk
Overview
This appears to be a code-review workflow skill made of markdown instructions, with no evidence of hidden code, persistence, credential use, or data exfiltration.
Install if you want a structured verification/code-review workflow. As with any review skill, check commands before approving test runs in sensitive repositories, but there is no artifact-backed reason to hold this skill for review.
Vulnerability Patterns
- 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
- Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Static analysis
No suspicious patterns detected.