The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill is designed to load and process a Google Service Account private key from a JSON file.
Evidence: scripts/gdocs_auth.py contains logic to open and parse a sensitive credential file to sign authentication tokens.
[EXTERNAL_DOWNLOADS]: The skill performs network requests to external Google API endpoints and requires the third-party cryptography Python library.
Evidence: scripts/gdocs_auth.py connects to oauth2.googleapis.com and docs.googleapis.com. It also imports multiple modules from the cryptography package.
[DATA_EXFILTRATION]: The authentication script explicitly disables SSL certificate verification for its network requests, which exposes sensitive authentication tokens and document content to potential man-in-the-middle attacks.
Evidence: scripts/gdocs_auth.py uses ssl._create_unverified_context() and passes it to urllib.request.urlopen calls.
[PROMPT_INJECTION]: The skill reads and processes data from external Google Documents, creating a surface for indirect prompt injection if those documents contain malicious instructions.
Ingestion points: scripts/gdocs_auth.py via the get_doc function.
Boundary markers: No specific delimiters or safety warnings are implemented in the provided patterns to distinguish between instructions and ingested content.
Capability inventory: The skill has the ability to write to documents and make outbound network connections.
Sanitization: There is no evidence of content sanitization or validation before the ingested data is processed by the agent.

gdocs-audit-report