microservices-patterns
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- [PROMPT_INJECTION] (SAFE): The skill instructions define strict grounding in reference files for architectural advice and code review, with no attempts to bypass safety filters or ignore system instructions.
- [EXTERNAL_DOWNLOADS] (SAFE): No remote scripts, binaries, or external packages are downloaded or referenced for execution.
- [COMMAND_EXECUTION] (SAFE): There are no subprocess calls, shell commands, or system-level interactions present in the skill files.
- [DATA_EXFILTRATION] (SAFE): No network operations or sensitive file path access (e.g., SSH keys, credentials) are performed.
- [INDIRECT_PROMPT_INJECTION] (LOW): The skill is designed to ingest and analyze user-provided code using architectural validation rules. This capability represents a surface for indirect injection if the processed data contains instructions, but the risk is mitigated as the skill lacks side-effect capabilities such as file modification or network egress.
Audit Metadata