The Agent Skills Directory

[COMMAND_EXECUTION]: The skill is built entirely around the execution of the obsidian CLI tool to perform file system operations (read, create, search) and application-level actions.
Evidence: Commands such as obsidian read, obsidian create, and obsidian search are used to interact with the user's local vault.
[DYNAMIC_EXECUTION]: The skill includes a command that allows for the execution of arbitrary JavaScript within the context of the Obsidian application, which can access the full Obsidian API and vault contents.
Evidence: The command obsidian eval code="..." is documented for running JavaScript in the app context, such as app.vault.getFiles().length.
[INDIRECT_PROMPT_INJECTION]: The skill creates an attack surface for indirect prompt injection because it is designed to read and process content from an Obsidian vault, which may contain untrusted data.
Ingestion points: The obsidian read, obsidian search, and obsidian daily:read commands ingest external data into the agent's context.
Boundary markers: None specified in the instructions to separate untrusted vault content from system instructions.
Capability inventory: The skill can write files (create, append), execute shell commands (obsidian ...), and run arbitrary JavaScript (eval).
Sanitization: There is no evidence of sanitization or validation of the content read from the vault before the agent processes it.

obsidian-cli