notion-knowledge-capture
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest untrusted data from conversation histories and transform it into documentation. It does not provide the agent with instructions to differentiate between conversational data and potential commands embedded within that data.
- Ingestion points: Conversation context and chat history are processed in Step 1 of the workflow defined in
SKILL.md. - Boundary markers: The skill lacks specified boundary markers (such as XML tags or delimiters) or explicit 'ignore previous instructions' warnings when the agent is analyzing the chat context.
- Capability inventory: The skill utilizes powerful modification capabilities, including
Notion:notion-create-pagesandNotion:notion-update-page, allowing it to create new content and modify existing indices or hub pages in a Notion workspace. - Sanitization: There are no instructions for the agent to sanitize or validate extracted content before it is committed to the Notion workspace, which could allow a malicious user to manipulate the resulting documentation or the agent's tool calls.
Audit Metadata