google-drive-upload
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEDATA_EXFILTRATIONCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill reads the user's configuration file at
~/.cowork-gdrive-config.jsonand transmits its contents (API key and script URL) along with Base64-encoded file data to an external Google Apps Script URL viacurl.- [COMMAND_EXECUTION]: The skill utilizes shell commands includingcatto read local configuration,base64andtrto process file data, andcurlto perform network operations.- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of arbitrary local files for processing and upload, which could contain malicious instructions designed to influence the agent's behavior. - Ingestion points: Local file paths provided by the user or generated by workflows (SKILL.md).
- Boundary markers: The skill does not implement delimiters or specific instructions to ignore embedded content within the files being processed.
- Capability inventory: File system read access (
cat), data transformation (base64), and network egress (curl). - Sanitization: There is no evidence of content sanitization or validation of the files before they are encoded and transmitted.
Audit Metadata