The Agent Skills Directory

[DATA_EXFILTRATION]: The skill accesses Google OAuth credentials and tokens stored in ~/.claude/.google/client_secret.json and ~/.claude/.google/token.json to authenticate requests to well-known Google API endpoints. This behavior is a core requirement for the skill's functionality and is documented neutrally. \n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface. 1. Ingestion points: Untrusted document content is read in scripts/docs_manager.rb via the read and structure commands. 2. Boundary markers: No delimiters or instructions are used to separate document content from agent instructions. 3. Capability inventory: The skill possesses extensive document modification permissions, including delete_content, replace_text, and insert_text in scripts/docs_manager.rb. 4. Sanitization: No filtering or escaping is applied to the ingested document text.

google-docs