performance-workload-modeling
Pass
Audited by Gen Agent Trust Hub on Oct 31, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied traffic and business data. It provides strong mitigation by requiring agents to verify input credibility and maintain strict separation between confirmed evidence and assumptions using mandatory labels.
- [SAFE]: The skill's scripts are restricted to deterministic output validation for automated testing. They utilize Python's standard library to parse Markdown tables and verify numeric calculations without external connectivity.
- [SAFE]: The documentation and prompts include explicit directives for least-privilege operations and data masking, ensuring that performance modeling tasks are handled with security and privacy in mind.
Audit Metadata