The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill exposes an obsidian eval command which allows the agent to execute arbitrary JavaScript code within the context of the running Obsidian application. While intended for developer use cases such as plugin debugging, this functionality represents a dynamic execution surface that could be leveraged to perform unauthorized actions within the host application environment.
[INDIRECT_PROMPT_INJECTION]: The skill enables the ingestion of untrusted data from an external vault, which creates a vulnerability to instructions embedded in that data.
Ingestion points: The skill reads vault file content (obsidian read, obsidian search), inspects the application's DOM (obsidian dev:dom), and retrieves console output (obsidian dev:console) as seen in SKILL.md.
Boundary markers: The instructions lack delimiters or explicit warnings to the agent to ignore instructions found within the vault data.
Capability inventory: The skill provides capabilities for arbitrary JavaScript execution (obsidian eval), file system writes (obsidian create, obsidian append), and taking screenshots (obsidian dev:screenshot) as described in SKILL.md.
Sanitization: There are no mentioned mechanisms for sanitizing or validating vault content before it is processed by the agent.

obsidian-cli