distributed-systems-patterns
Pass
Audited by Gen Agent Trust Hub on Oct 9, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill consists solely of informational Markdown files providing architectural patterns and decision frameworks. It contains no executable scripts, shell commands, or external dependencies.
- [INDIRECT_PROMPT_INJECTION]: The skill defines a process for the agent to ingest and analyze external repository data and user requirements. This creates a theoretical surface where malicious instructions in analyzed project files could attempt to influence the agent's architectural recommendations. Ingestion points occur where the agent searches for manifests, topics, and schemas (SKILL.md, references/event-driven-review.md). Boundary markers are implemented by requiring that inferred facts be prefixed with 'assumed:' in the output templates (assets/pattern-brief.md, assets/edm-decision-brief.md). The skill's capability inventory is empty, lacking functions for file writing, network access, or arbitrary command execution, which mitigates the risk of any injection.
Audit Metadata