Security audit
Agent Memory Tiers
Security checks for vulnerabilities and agentic risk
Overview
This skill is a disclosed documentation-only workflow for maintaining small local agent memory files, with no scripts, network access, or hidden execution.
Install this only for agents where cross-run memory is desired. Keep L0.md and L1.md concise, avoid storing secrets or unreviewed instructions in memory files, and periodically review the SOUL.md memory instructions because they influence future agent behavior.
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.