google-docs-skill

Pass

Audited by Gen Agent Trust Hub on Mar 29, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill allows the agent to ingest content from Google Docs and local files, which could contain instructions intended to influence agent behavior.\n
  • Ingestion points: read, from-markdown, and file-based create/update commands.\n
  • Boundary markers: None are defined to separate document content from agent instructions.\n
  • Capability inventory: The skill can export files to the local system, share documents with third parties, and upload images with public access.\n
  • Sanitization: Ingested text is not sanitized or escaped.\n- [DATA_EXPOSURE]: The script accesses authentication tokens and searches for credentials in other related skill directories to facilitate setup.\n- [EXTERNAL_DOWNLOADS]: Uses official Google API libraries to interact with Google services.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 29, 2026, 03:31 AM