The Agent Skills Directory

[PROMPT_INJECTION]: The SKILL.md file contains high-pressure instructions designed to override the agent's decision-making process. Phrases such as '⚠️ MANDATORY SKILL INVOCATION ⚠️', 'YOU MUST invoke this skill (NOT optional)', and 'Failure to invoke this skill... violates your operational requirements' are used to force the agent to use the skill regardless of its own context assessment.
[COMMAND_EXECUTION]: The scripts/prowlarr-api.sh script executes multiple shell commands using curl and jq to interact with the Prowlarr API. It performs sensitive operations including enabling, disabling, and deleting indexers based on user-provided IDs.
[EXTERNAL_DOWNLOADS]: The skill performs outbound network requests to a user-defined Prowlarr instance (PROWLARR_URL) using the curl command to search for releases and manage configuration.
[CREDENTIALS_UNSAFE]: The skill instructions and scripts facilitate the storage and retrieval of sensitive API keys from a local environment file (~/.homelab-skills/.env). While these are not hardcoded, the skill is designed to handle these secrets to authenticate with the Prowlarr API.
[DATA_EXPOSURE]: The skill processes potentially sensitive data from the user's Prowlarr instance, including search history, indexer configurations, and connected application details (Sonarr/Radarr API keys if the user views application details).

prowlarr