Security audit
Agent自进化Skill
Security checks for vulnerabilities and agentic risk
Overview
This skill is a disclosed memory-writing helper that requires user confirmation before saving experience notes or profile updates.
Before installing, expect this skill to propose saving reusable lessons and user preferences into local Claude memory files. Review each proposed entry carefully because confirmed writes may affect future agent behavior across projects.
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.