The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes various shell commands to perform diagnostics, including node for automated health checks, launchctl for managing system services, and find for file system audits.
[CREDENTIALS_UNSAFE]: The troubleshooting procedures instruct the agent to read integration tokens directly from local environment configuration files using grep to verify their validity.
[PERSISTENCE]: The skill interacts with the macOS launchd system to monitor and manage the state of background synchronization jobs stored in ~/Library/LaunchAgents/.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The agent reads and processes an auto-generated health report and samples content from user-owned markdown files in the notes vault.
Boundary markers: There are no explicit delimiters or instructions for the agent to ignore potentially malicious content embedded within the vault files or generated reports.
Capability inventory: The skill has access to shell execution (node, launchctl, curl), extensive file system read access, and network capabilities.
Sanitization: No sanitization or validation of the input data from the vault or report files is specified before the agent evaluates the content.

vault-health