The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install a third-party package (@mariozechner/gdcli) from the global NPM registry. This package is not maintained by a verified organization or a well-known service, posing a supply chain risk.
[CREDENTIALS_UNSAFE]: The skill explicitly documents and manages sensitive files containing OAuth client secrets and user access tokens stored in the ~/.gdcli/ directory. Providing an agent with instructions to access and manipulate these credential files increases the risk of accidental exposure or misuse of Google Drive permissions.
[INDIRECT_PROMPT_INJECTION]: The skill enables the agent to search and download file content from an external source (Google Drive) that could contain malicious instructions.
Ingestion points: File content and metadata retrieved via gdcli search, gdcli download, and gdcli ls (SKILL.md).
Boundary markers: None present to distinguish data from instructions.
Capability inventory: Subprocess calls to the gdcli tool for listing, searching, downloading, and sharing files (SKILL.md).
Sanitization: No sanitization or validation of the retrieved file content is specified.

gdcli