llm-wiki-retrieval-architect
Pass
Audited by Gen Agent Trust Hub on Aug 11, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill serves exclusively as an architectural guide and procedural framework. It instructs the agent on how to evaluate retrieval failures and select appropriate technology tiers (lexical, semantic, hybrid, or graph-based) without executing any high-risk operations.
- [SAFE]: Includes comprehensive security and privacy controls in both the procedural instructions and reference documentation. It explicitly mandates PII/secret scanning and query-time metadata filtering for tenant isolation and sensitivity boundaries.
- [EXTERNAL_DOWNLOADS]: The reference documentation contains links to official documentation for well-known technologies and trusted organizations, such as Microsoft, Anthropic, SQLite, and established open-source projects like Qdrant, Weaviate, and LlamaIndex. These are used for architectural research and do not involve untrusted remote code execution.
Audit Metadata