notion-hardened
Pass
Audited by Gen Agent Trust Hub on Apr 21, 2026
Risk Level: SAFE
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill manages its authorization by reading an API key from a dedicated local configuration file (~/.config/notion/api_key). This is a standard and isolated method for providing credentials to the integration.
- [DATA_EXFILTRATION]: All network operations are directed to api.notion.com, which is the official and trusted domain for the Notion API service. No unauthorized data transmission was detected.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external Notion pages and databases, which could potentially contain malicious instructions. To mitigate this risk, the skill includes explicit 'Security Guardrails' that instruct the agent to limit its search scope, confirm destructive operations, and sanitize data before external sharing.
- Ingestion points: Notion API responses containing page and database content.
- Boundary markers: Present as explicit behavioral instructions in the 'Security Guardrails' section of the SKILL.md file.
- Capability inventory: Network requests and data manipulation via curl commands.
- Sanitization: Implemented through high-level instructions to the agent rather than programmatic filtering of the API responses.
Audit Metadata