The Agent Skills Directory

[DATA_EXFILTRATION]: The skill reads from a sensitive file path: ~/Library/Application Support/amaran Desktop/*/amaran.db. This SQLite database contains mesh network encryption keys (net_key, app_key) and device addresses. Accessing another application's internal database to extract credentials is a sensitive operation, although it is the core mechanism for this skill's functionality.- [COMMAND_EXECUTION]: The skill instructs the agent to install a local Python package and execute the amaran CLI tool. This allows the agent to perform hardware discovery and direct control of studio lights via Bluetooth commands built within the Python environment.- [PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection through its data processing pipeline. * Ingestion points: Reads device names and network metadata from the local amaran.db file and processes user-provided target substrings via CLI arguments. * Boundary markers: The instructions do not provide explicit delimiters or warnings to separate data retrieved from the database or CLI output from the agent's internal reasoning. * Capability inventory: The skill can interact with the local filesystem (read/write config), query an external SQLite database, and communicate over Bluetooth (BLE Mesh) to modify hardware states. * Sanitization: Lighting parameters are validated against numeric ranges, and device identifiers are sanitized using a regex-based slugging function to remove special characters.

amaran-lights