The Agent Skills Directory

[COMMAND_EXECUTION]: The skill makes extensive use of system commands and scripting for local monitoring.
The shell scripts and TypeScript implementation use osascript (AppleScript/JXA) to poll Apple Notes for metadata.
It uses pgrep to monitor the status of iCloud system daemons (bird, cloudd).
These operations are restricted to the local environment and are essential for the skill's stated purpose of observability.
[PERSISTENCE]: The documentation explicitly instructs users to deploy the health check script as a macOS launchd agent.
This mechanism allows the script to run periodically (e.g., every 5 minutes) across sessions.
While this is a persistence mechanism, it is used here for legitimate monitoring purposes and follows standard macOS administration patterns.
[INDIRECT_PROMPT_INJECTION]: The skill has a data ingestion surface through local command outputs.
Ingestion points: System output from osascript (note/folder counts) and pgrep are processed in src/observability/metrics.ts and the bash scripts.
Boundary markers: The TypeScript implementation uses JSON.parse() for structured data ingestion from shell outputs, reducing the risk of injection.
Capability inventory: The skill uses execSync for shell command execution and appendFileSync for logging to the file system.
Sanitization: The instructions explicitly mandate sanitizing shell output and prohibit the interpolation of sensitive content like note titles, bodies, or account names into notifications or telemetry.

apple-notes-observability