The Agent Skills Directory

[COMMAND_EXECUTION]: The skill features the obsidian eval command, which allows for the execution of arbitrary JavaScript code strings within the Obsidian application context. While documented for plugin development, this provides a vector for dynamic code execution.
[PROMPT_INJECTION]: The skill handles untrusted data from Obsidian vault files, creating a surface for indirect prompt injection.
Ingestion points: Data enters the agent context through file-reading commands such as obsidian read, obsidian search, and obsidian properties.
Boundary markers: The instructions do not define delimiters or specific constraints to prevent the agent from interpreting note content as instructions.
Capability inventory: The skill includes extensive file modification capabilities (create, append, move, delete) and JavaScript execution (eval) that could be leveraged by malicious content in a vault.
Sanitization: There is no evidence of content sanitization or validation for the data retrieved from the vault files.
[SAFE]: The skill references documentation hosted on the official GitHub repository for Obsidian (obsidianmd), which is a well-known and trusted source.

obsidian-cli