The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes an obsidian command-line tool to perform vault operations and development tasks.
[COMMAND_EXECUTION]: The obsidian eval command allows for the execution of arbitrary JavaScript code within the context of the running Obsidian application, which can be used to bypass app-level security or access internal data.
[DATA_EXFILTRATION]: The skill provides the agent with the ability to read sensitive local data, including the full content of any note in the vault, search results, and application console logs, through commands like obsidian read, obsidian search, and obsidian dev:console.
[EXTERNAL_DOWNLOADS]: The skill references documentation hosted on the official Obsidian domain (https://help.obsidian.md/cli).
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it processes untrusted data from local notes while maintaining high-privilege capabilities.
Ingestion points: vault content is ingested via obsidian read, obsidian search, and obsidian daily:read commands.
Boundary markers: No delimiters or safety instructions are defined to help the agent distinguish between its instructions and the content retrieved from notes.
Capability inventory: The skill can modify the local file system (create, append, property:set) and execute dynamic code (eval).
Sanitization: There is no evidence of sanitization or filtering applied to note content before it is processed by the agent.

obsidian-cli