client-package-notion
Pass
Audited by Gen Agent Trust Hub on Mar 31, 2026
Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill is designed to read strategy documents, lead lists, and campaign assets from the local filesystem (
clients/directory) and transmit that data to external cloud services (Notion and Google Sheets). While these are well-known services, the process involves moving sensitive client data from a local environment to the cloud. - [PROMPT_INJECTION]: The skill exhibits a vulnerability to indirect prompt injection because it processes untrusted content from markdown and CSV files without sanitization or boundary markers.
- Ingestion points: Reads various file types (Markdown, CSV) from
clients/<client_name>/including strategies, campaign assets, and lead lists. - Boundary markers: Absent. The skill instructions direct the agent to "Read the full .md content" and integrate it directly into the output structure.
- Capability inventory: Creating and updating content in Notion via API, creating Google Sheets via the Rube MCP server, and potentially triggering automated outreach through the
setup-outreach-campaignskill. - Sanitization: No mechanisms are described to sanitize the ingested text for embedded instructions that could hijack the agent's logic or trigger unauthorized actions in connected skills.
Audit Metadata