protection-shield
Pass
Audited by Gen Agent Trust Hub on Nov 13, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from various sources which establishes a surface for indirect prompt injection attacks.
- Ingestion points: Processes system logs (via 'journalctl' and 'Get-WinEvent'), configuration hashes, and external security advisories/vendor hints (coordinated through 'security-tech-stack-check').
- Boundary markers: The instructions do not define specific delimiters or isolation markers for the ingested external content.
- Capability inventory: The skill has high-impact capabilities including stopping services, modifying firewall rules, rotating credentials, and executing security patches.
- Sanitization: The instructions include a requirement to redact secrets from captured logs and configurations before further processing.
- [COMMAND_EXECUTION]: The skill utilizes system-level commands and external scripts to perform its intended defensive functions.
- Evidence: The skill executes native system administration tools (e.g., firewall managers, service managers) and invokes a Python comparison script ('scripts/techshot.py') located in the 'tech-detect' skill directory. These operations are scoped to authorized assets and incident response workflows.
Audit Metadata