The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill functions as a standard utility for managing Google Keep notes through a verified third-party library.
[CREDENTIALS_UNSAFE]: The skill manages authentication tokens locally. It stores a master token in ~/.config/gkeep/token.json after a successful login. The implementation in gkeep.py correctly applies chmod 0o600 to the file, ensuring that sensitive data is only accessible by the owner, which aligns with security best practices for credential management.
[PROMPT_INJECTION]: The skill retrieves untrusted data from an external source (Google Keep), creating a surface for indirect prompt injection.
Ingestion points: Note titles and bodies are fetched via cmd_list, cmd_get, and cmd_search in gkeep.py.
Boundary markers: The content is printed directly to the output without delimiters or instructions to ignore embedded commands.
Capability inventory: The skill's capabilities are limited to managing Google Keep notes; no arbitrary command execution or filesystem access (beyond the token file) was detected.
Sanitization: No sanitization or escaping is performed on the retrieved note content.

gkeep