The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill reads untrusted markdown files provided via the source_markdown_path, creating an attack surface for instructions embedded within the processed data.
Ingestion points: SKILL.md (reads external markdown files into the agent's context).
Boundary markers: Absent; the instructions do not define delimiters or warnings to ignore instructions inside the imported files.
Capability inventory: The skill performs file system reads, file writes to the Obsidian vault, and execution of subsequent formatting commands.
Sanitization: Absent; there is no evidence of filtering or validation of the markdown content before it is processed or used to trigger other actions.
[COMMAND_EXECUTION]: The skill's workflow includes running a command named format-note. Executing secondary logic or scripts on imported, untrusted data can lead to unintended command execution if the content is not properly isolated.

create-note