The Agent Skills Directory

[COMMAND_EXECUTION]: The skill requires the curl binary to be present on the system path and uses it to perform network requests and API interactions.
[EXTERNAL_DOWNLOADS]: The skill is designed to fetch live data from remote services and servers, specifically targeting endpoints like Home Assistant's /api/states via a user-defined HA_URL.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to how it processes external data.
Ingestion points: External data enters the agent context via network requests to remote system status endpoints and Home Assistant entity state APIs in SKILL.md (Procedure step 6).
Boundary markers: Absent. The instructions do not specify the use of delimiters or provide warnings to ignore instructions that might be embedded within the retrieved remote content.
Capability inventory: The skill possesses network capabilities via curl as defined in the Prerequisites and Procedure sections.
Sanitization: Absent. There is no requirement or logic provided to sanitize, filter, or validate the content fetched from remote systems before it is presented to the user or processed by the agent.

scrying