google-docs
Pass
Audited by Gen Agent Trust Hub on Mar 8, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The script
create.shinteracts with well-known Google services atgoogleapis.comto exchange credentials for access tokens and perform document operations.\n- [COMMAND_EXECUTION]: The skill executes local system commands includingopensslfor generating RS256 signatures andjqfor handling JSON data structure construction in thecreate.shscript.\n- [PROMPT_INJECTION]: Indirect prompt injection surface detected:\n - Ingestion points: The
CONTENTargument increate.shaccepts arbitrary text from the agent context to be written to a document.\n - Boundary markers: No specific delimiters or instructions are used to separate user-provided data from the document structure or to warn the agent about embedded content.\n
- Capability inventory: The skill possesses the capability to create and modify Google Documents via authenticated API calls using
curl.\n - Sanitization: While
jq --argis used to prevent shell or JSON structure injection, the text content itself is not sanitized for potential instructions that could influence an agent reading the document later.
Audit Metadata