notion-migration-deep-dive

Pass

Audited by Gen Agent Trust Hub on Mar 25, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns, obfuscation, or unauthorized data exfiltration techniques were detected in the skill instructions or code snippets.
  • [DATA_EXPOSURE]: The skill correctly implements security best practices by recommending the use of environment variables (NOTION_TOKEN) for API authentication rather than hardcoding secrets.
  • [INDIRECT_PROMPT_INJECTION]: The skill functions as an ETL tool, ingesting data from external files (CSV) and Notion databases. This represents an ingestion surface for indirect prompt injection if the source data contains malicious instructions. However, the risk is inherent to the data migration use case and is mitigated by property-type mapping and length validation.
  • Ingestion points: importFromCSV (local files), exportDatabase (Notion API).
  • Boundary markers: Absent.
  • Capability inventory: Bash(npm/node/kubectl), Write (file system).
  • Sanitization: Implements property type validation and slice(0, 2000) for rich text content.
  • [EXTERNAL_DOWNLOADS]: Dependencies are restricted to reputable, official libraries from standard package registries (NPM and PyPI).
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 25, 2026, 04:41 PM