The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill retrieves and processes user-controlled content from Apple Notes via the read-note and search-notes commands.
Ingestion points: The contents of macOS Notes are read into the agent's context through scripts/notes.sh using AppleScript.
Boundary markers: The instructions do not define explicit delimiters or instructions to help the agent distinguish between note content and its system instructions.
Capability inventory: The skill possesses the ability to execute local scripts (osascript, python3), manage system processes (pgrep, open), and write to a local log file (logs/notes.log).
Sanitization: While the script uses html.escape when writing notes to ensure valid HTML, content retrieved for reading is passed as plaintext to the agent without additional filtering for potential injection instructions.
[COMMAND_EXECUTION]: The skill relies on scripts/notes.sh to execute local macOS commands. The script follows security best practices by passing data through environment variables and using the argv interface for AppleScript, which prevents common shell injection vulnerabilities during command construction.

macos-notes