The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides an obsidian eval command in the references/cli-reference.md file, which enables the execution of arbitrary JavaScript code within the application's environment. This grants the agent direct access to the internal Obsidian API and the underlying system via the Electron framework.- [EXTERNAL_DOWNLOADS]: The obsidian plugin:install command allows the agent to fetch and install third-party community plugins. This mechanism facilitates the execution of unverified remote code with the same privileges as the main application.- [COMMAND_EXECUTION]: The skill defines several high-risk CLI commands, including obsidian command for executing internal app functions and obsidian dev:screenshot for capturing images of the vault content, which could lead to unauthorized data exposure.- [PROMPT_INJECTION]: The skill exposes a significant surface for indirect prompt injection. Ingestion points: The agent can ingest untrusted data via obsidian read, obsidian search, obsidian dev:dom, and obsidian web. Boundary markers: No boundary markers or 'ignore instructions' delimiters are specified for processing this data. Capability inventory: The skill possesses high-impact capabilities including obsidian eval (arbitrary JS), obsidian plugin:install (RCE), and file system modifications (create, delete, move). Sanitization: No sanitization or validation logic is present to filter malicious instructions from processed content.

obsidian