mcp-media-inbox-triage

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it processes untrusted data from a shared Gmail inbox to perform classification and routing.
  • Ingestion points: The skill reads the sender name, subject, and plain-text body of emails from Gmail as defined in SKILL.md and SETUP.md.
  • Boundary markers: The classification prompt in SKILL.md lacks explicit structural delimiters (like XML tags or triple quotes) to separate the untrusted email content from the agent's instructions.
  • Capability inventory: The skill possesses capabilities to write to Slack, Google Sheets, and create Gmail drafts via Zapier MCP as outlined in SKILL.md and SETUP.md.
  • Sanitization: While the skill instructs the AI to summarize rather than expose full email bodies, it does not include programmatic sanitization or escaping of the ingested text.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 06:14 PM
Security Audit — agent-trust-hub — mcp-media-inbox-triage