The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill metadata specifies a Go installation command for github.com/steipete/sonoscli/cmd/sonos@latest. This downloads and executes code from a third-party repository not included in the trusted provider list.
[COMMAND_EXECUTION]: The skill relies on executing the sonos CLI tool to perform network discovery, status checks, and speaker control. This grants the agent the ability to execute shell commands with user-supplied arguments (e.g., --name, --ip, and search queries).
[CREDENTIALS_UNSAFE]: The documentation notes that optional features require SPOTIFY_CLIENT_ID and SPOTIFY_CLIENT_SECRET. While not hardcoded, the presence of these secrets in the environment represents a high-value target for exfiltration.
[PROMPT_INJECTION]: The skill processes external data from Spotify search results (sonos smapi search) and local speaker queues (sonos queue list).
Ingestion points: Data enters the agent context via output from Spotify search queries and local device queue listings.
Boundary markers: None identified in the provided instructions; untrusted data is likely interpolated directly into the agent's context.
Capability inventory: The skill has the capability to execute shell commands via the sonos binary and access network resources.
Sanitization: There is no evidence of sanitization or filtering applied to the data retrieved from external APIs or local network devices before it is processed by the agent.

sonoscli