The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied descriptions of systems or processes to generate suspension plans. This creates an ingestion surface for untrusted data.
Ingestion points: The skill accepts user input via the command template /sleep design a graceful suspension plan for this [system/process/notification channel].
Boundary markers: Absent; the skill does not explicitly define delimiters for the user-provided system description.
Capability inventory: The skill identifies curl as a required binary on the system PATH for interacting with the Home Assistant API.
Sanitization: Absent; the skill does not specify filtering or escaping for the user-supplied identifiers before plan generation.
[COMMAND_EXECUTION]: The skill identifies curl as a required binary for interacting with the Home Assistant environment. This is a standard utility for the intended integration and is used according to documented requirements.
[CREDENTIALS_UNSAFE]: The skill references the requirement for HA_TOKEN and HA_URL. It follows best practices by instructing that these credentials be managed through environment variables rather than being hardcoded in the skill's instructions or scripts.

sleep