NYC
skills/smithery/ai/himalaya/Gen Agent Trust Hub

himalaya

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS] (SAFE): The skill facilitates the installation of the 'himalaya' binary using Homebrew (brew install himalaya). While the source organization (pimalaya) is not on the predefined trusted list, Homebrew is a standard, reputable package manager for this type of utility.
  • [DATA_EXFILTRATION] (SAFE): The skill's primary purpose is email management, which involves reading and sending messages. The analysis found no evidence of unauthorized data transfer to third-party domains. All network operations are directed toward the user's configured IMAP/SMTP servers.
  • [CREDENTIALS_UNSAFE] (SAFE): The configuration example correctly avoids hardcoding secrets, instead recommending the use of secure command-based password retrieval (e.g., pass show email/imap).
  • [INDIRECT_PROMPT_INJECTION] (LOW):
  • Ingestion points: The agent can ingest untrusted data through commands like himalaya message read and himalaya envelope list which display content from external senders.
  • Boundary markers: Absent. The instructions do not provide explicit delimiters or warnings to the agent regarding the processing of untrusted email content.
  • Capability inventory: The agent has the ability to read, list, move, delete, and send emails.
  • Sanitization: None specified in the markdown instructions. As this is an inherent risk of email-processing skills, it is rated as LOW severity.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 05:48 PM