The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads arbitrary content from Google Sheets, creating a vector for malicious instructions to influence agent behavior.
Ingestion points: Data is ingested from spreadsheets in scripts/sheets.py via the get-text and get-range functions.
Boundary markers: The skill does not use specific delimiters or protective instructions when presenting sheet data to the agent.
Capability inventory: The skill has extensive write permissions, including update-range, append-rows, and batch-update in scripts/sheets.py, which could be abused if the agent follows injected instructions.
Sanitization: Cell content is retrieved and displayed without sanitization or validation against instruction patterns.
[EXTERNAL_DOWNLOADS]: The skill interacts with an external service to facilitate the OAuth authentication flow and refresh tokens.
Evidence: scripts/auth.py communicates with https://google-workspace-extension.geminicli.com to exchange authorization codes and refresh access tokens.
Context: While this is a functional part of the skill's authentication mechanism, it involves sending refresh tokens to a third-party endpoint.

google-sheets