The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill documentation includes an installation command that downloads a script from a remote URL and pipes it directly to the shell for execution. Evidence: curl -sSL https://raw.githubusercontent.com/pimalaya/himalaya/master/install.sh | PREFIX=~/.local sh in SKILL.md.
[EXTERNAL_DOWNLOADS]: The skill references and downloads code from a repository that is not part of the trusted vendors list. Evidence: https://raw.githubusercontent.com/pimalaya/himalaya/master/install.sh.
[INDIRECT_PROMPT_INJECTION]: The skill enables the agent to read and process emails from external sources, which can contain malicious instructions designed to manipulate the agent. Ingestion points: himalaya message read and himalaya envelope list commands in SKILL.md ingest untrusted content into the agent's context. Boundary markers: Absent. There are no instructions or delimiters to help the agent distinguish between its own instructions and the content of the emails it reads. Capability inventory: The agent has access to terminal commands, file system operations, and email sending capabilities (himalaya template send). Sanitization: Absent. No mention of sanitizing or filtering email content before it is processed.

himalaya