notion-reliability-patterns
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill correctly handles authentication by reading the
NOTION_TOKENfrom environment variables (process.envin TypeScript andos.environin Python), avoiding the risk of hardcoded credentials. - [SAFE]: All external dependencies are official or widely-trusted libraries, specifically
@notionhq/client,notion-client, andlru-cache. - [INDIRECT_PROMPT_INJECTION]: The patterns facilitate the ingestion of data from the Notion API, which constitutes a potential attack surface for indirect prompt injection. However, given the context of a developer utility for reliability, this is an inherent characteristic rather than a flaw.
- Ingestion points: Notion API retrieval methods in
references/implementation.md. - Boundary markers: Not implemented in the skeleton code.
- Capability inventory: The skill environment allows
Read,Write, andEdittool usage. - Sanitization: Not present in the provided boilerplate; users should implement content sanitization based on their specific application needs.
Audit Metadata