notion-artifact-capture
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFE
Full Analysis
- [UNVERIFIABLE_DEPENDENCIES_AND_REMOTE_CODE_EXECUTION]: The skill does not perform any remote code execution or installation of untrusted packages. It uses standard Python library modules (urllib) for network communication and relies on locally cloned scripts.
- [DATA_EXPOSURE_AND_EXFILTRATION]: Network activity is restricted to the official Notion API domain (api.notion.com). Notion integration tokens are managed securely through environment variables and are never hardcoded or printed in diagnostic output.
- [PRIVILEGE_ESCALATION]: The skill does not attempt to gain elevated privileges. The bootstrap script explicitly restricts file permissions on its generated configuration file to the current user only (read/write), which is a proactive security measure.
- [INDIRECT_PROMPT_INJECTION]: The skill represents a data ingestion surface as it processes local Markdown files to save as Notion page bodies. Evidence chain: 1. Ingestion: save_artifact.py reads from a file path provided via the --markdown-file argument. 2. Boundary markers: Absent. 3. Capability inventory: The script writes content to the user's Notion workspace via the API. 4. Sanitization: Absent, as original formatting must be preserved for the Notion page content. This surface is a functional requirement for the skill's purpose and is controlled by user-supplied arguments.
Audit Metadata