gmail-label

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection due to its handling of untrusted email content during the classification process.
  • Ingestion points: The scripts/gmail_label_fetch.py script retrieves email subjects and snippets from the Gmail API and stores them in a JSON file for processing.
  • Boundary markers: The subagent prompt described in SKILL.md lacks delimiters or specific instructions to ignore embedded commands within the ingested email snippets.
  • Capability inventory: The skill possesses the capability to modify Gmail account data by creating and applying labels via scripts/gmail_label_apply.py.
  • Sanitization: There is no evidence of sanitization or filtering of the email metadata before it is used for AI-driven classification.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 02:18 PM