skills/glebis/claude-skills/gmail/Gen Agent Trust Hub

gmail

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFEDATA_EXFILTRATIONPROMPT_INJECTIONNO_CODEEXTERNAL_DOWNLOADS
Full Analysis
  • [DATA_EXFILTRATION] (LOW): The skill explicitly directs the storage and access of sensitive OAuth credentials and authentication tokens in the ~/.gmail_credentials/ directory. Accessing and managing these secrets is a sensitive data operation.
  • [PROMPT_INJECTION] (LOW): The skill is vulnerable to Indirect Prompt Injection. 1. Ingestion points: Fetches untrusted data from email subjects, snippets, and bodies (via scripts/gmail_search.py). 2. Boundary markers: Uses Markdown headers and blockquotes for output, but lacks explicit delimiters or warnings to ignore instructions embedded in email content. 3. Capability inventory: The skill can read emails, download attachments to the local filesystem, and modify email labels. 4. Sanitization: No sanitization or filtering of email content is documented.
  • [NO_CODE] (SAFE): The logic file scripts/gmail_search.py is referenced but not provided in the skill package, limiting the audit to instructional content.
  • [EXTERNAL_DOWNLOADS] (SAFE): The skill requires standard Google API libraries which are considered trusted sources.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:12 PM