The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it reads and processes external data from the user's Obsidian vault.
Ingestion points: Note content is retrieved from /Users/fredrik/Library/Mobile Documents/iCloud~md~obsidian/Documents/personal using rg, fd, and grep.
Boundary markers: The instructions do not specify any delimiters (e.g., XML tags) or safety warnings to prevent the agent from following instructions found within the note content.
Capability inventory: The skill allows the agent to execute shell commands (fd, rg, grep, find, date), create new markdown files, and modify existing files within the vault.
Sanitization: No validation or sanitization of the retrieved note content is performed before the agent processes it.
[COMMAND_EXECUTION]: The skill utilizes shell utilities like rg (ripgrep), fd, grep, and find to search and interact with the markdown files on the local filesystem. This is consistent with the skill's purpose as an Obsidian vault manager.

obsidian