The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses system shell commands, including obsidian, open, and sleep, to automate the Obsidian application and perform file operations.
[DATA_EXFILTRATION]: The skill interacts with the local file system to read vault content and uses CLI commands to reveal the absolute system path of the Obsidian vault.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes content from user-generated notes and external web sources.
Ingestion points: SKILL.md (uses obsidian read and obsidian search to fetch data).
Boundary markers: Absent. No specific delimiters are defined for the ingested content.
Capability inventory: SKILL.md (has the ability to create, append, and move files, and update note properties).
Sanitization: Partial. The skill instructs the agent to avoid notes with privacy tags (#Private, #Key) but does not validate the content of the notes it processes.

obsidian-knowledge