Security audit
output-validation
Security checks for vulnerabilities and agentic risk
Overview
This is a simple local validation guide for checking output files and does not show suspicious behavior.
Reasonable to install as an instruction-only local validator. Use it only on files you intentionally provide, and expect the example Python checks to raise assertions on malformed outputs rather than fixing them automatically.
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.