The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill exposes the obsidian eval command, which allows the agent to execute arbitrary JavaScript code strings within the Obsidian application instance. While intended for developer use, this capability provides a powerful vector for local code execution.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill facilitates access to potentially sensitive user data stored within Obsidian vaults through several commands:
obsidian read and obsidian search allow the agent to extract the contents of any note or file.
obsidian dev:screenshot and obsidian dev:dom allow the agent to capture visual or textual representations of the application's internal state.
obsidian dev:console permits reading application logs which may contain sensitive debugging information.
[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection attacks where malicious instructions stored in vault files could influence the agent's behavior.
Ingestion points: Vault data is ingested via obsidian read, obsidian search, obsidian dev:dom, and obsidian dev:console (SKILL.md).
Boundary markers: None identified; there are no instructions for the agent to use delimiters or ignore embedded instructions when processing vault content.
Capability inventory: The skill possesses write and execute capabilities including obsidian create, obsidian append, obsidian property:set, and obsidian eval (SKILL.md).
Sanitization: There are no documented mechanisms for sanitizing or validating the content retrieved from the vault before it is processed by the agent.

obsidian-cli