notion-knowledge-capture
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- Indirect Prompt Injection Surface: The skill is designed to process user conversation history and notes, which are untrusted external inputs, to generate structured Notion pages. \n
- Ingestion points: The workflow in
SKILL.mdpulls context directly from chat sessions to extract facts and decisions. \n - Boundary markers: There are no explicit instructions or delimiters defined to help the agent distinguish between intended content and potentially malicious instructions embedded within the conversation data. \n
- Capability inventory: The skill possesses write access to external workspaces via tools such as
Notion:notion-create-pagesandNotion:notion-update-page. \n - Sanitization: Content extraction relies on model interpretation without structured validation or sanitization of the input's intent. \n- Data Flow to External Systems: The skill facilitates the transfer of processed information into a persistent team knowledge base in Notion. This creates a path where unverified data from a conversation could be permanently stored and accessed by others in the organization.
Audit Metadata