parcel-tracking
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface because it processes untrusted data received from external email senders.
- Ingestion points: The skill searches and reads the content of recent Gmail dispatch and shipping emails in
SKILL.md(Step 1). - Boundary markers: There are no specific boundary markers or delimiters specified to isolate untrusted email text from the agent's core instructions.
- Capability inventory: The skill can invoke web scraping or browser automation tools (e.g., Chrome MCP, Playwright) to navigate and scrape data from courier web pages (Step 3).
- Sanitization: No input sanitization or filtering is prescribed for the text extracted from email bodies before it is used to interpolate fields or display content.
Audit Metadata