email-management-expert

Warn

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • EXTERNAL_DOWNLOADS (MEDIUM): The README instructions recommend downloading a ZIP file from an untrusted source (github.com/patrickfreyer/apple-mail-mcp/releases) and unzipping it into a sensitive directory (~/.claude/skills/). This allows for the introduction of unverified code from an author not in the trusted scope.
  • PROMPT_INJECTION (LOW): The skill is vulnerable to Indirect Prompt Injection (Category 8). 1. Ingestion points: Untrusted content from incoming emails accessed via search_emails and get_recent_emails (referenced in examples/inbox-zero-workflow.md). 2. Boundary markers: Absent. No instructions are given to the agent to treat email content as untrusted data or ignore instructions within email bodies. 3. Capability inventory: High-impact tools including forward_email, manage_trash, reply_to_email, export_emails, and manage_drafts (referenced in examples/inbox-zero-workflow.md). 4. Sanitization: None. No escaping or validation is performed on external email content.
  • COMMAND_EXECUTION (LOW): The skill documentation includes shell commands for installation (unzip, cp) and describes workflows using tools that perform file system and network operations related to email management.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 17, 2026, 04:31 PM