The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes Python scripts that interact with the host system using the subprocess module. Scripts such as scripts/check_ob_installed.py, scripts/validate_vault_config.py, and scripts/generate_systemd_service.py execute commands like node --version and ob sync-status. These interactions use argument lists to avoid shell injection vulnerabilities.
[CREDENTIALS_UNSAFE]: The skill documents the authentication process for Obsidian Sync, which involves passing email and password information via CLI flags or environment variables. It explicitly notes that the ob tool stores these credentials in flat files within the vault's .obsidian-sync/ directory, which is standard behavior for the supported CLI.
[PERSISTENCE]: The skill provides the scripts/generate_systemd_service.py script to automate the creation of systemd unit files. This allows the ob sync --continuous process to run as a background daemon and restart automatically, which is a persistence mechanism intended for server-side automation.
[INDIRECT_PROMPT_INJECTION]: The skill processes vault data which could potentially contain malicious instructions from untrusted sources. Ingestion points: Content is ingested from local files and remote Obsidian Sync vaults. Boundary markers: No specific delimiters are used to wrap external content within the skill's logic. Capability inventory: The skill can perform filesystem writes, network synchronization via the ob tool, and service management. Sanitization: The skill relies on the underlying CLI tool's data handling and uses Path.resolve() in its Python scripts for path management.

obsidian-headless