The Agent Skills Directory

[SAFE]: The skill uses environment variables (~/.claude-homelab/.env) to handle sensitive API keys and URLs safely, following standard secret management practices.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from external APIs (TV show search descriptions from Sonarr/TVDB).
Ingestion points: Data enters the context through scripts/sonarr.sh via the search and search-json subcommands when fetching series metadata.
Boundary markers: No specific LLM boundaries are enforced around the fetched metadata text.
Capability inventory: Capabilities are limited to standard HTTP REST calls via curl to a Sonarr API instance to manage TV series libraries.
Sanitization: The script securely sanitizes user inputs by utilizing jq -sRr @uri for URL encoding and jq --arg for JSON payload construction, safely preventing shell or command injection.

sonarr