List Available Guardrails

Retrieve a list of all available guardrails for content validation. **Note**: This endpoint returns a plain array of guardrails, not a paginated response. All available guardrails are returned in a single request. Rightbrain currently ships two built-in guardrails, both applied to the input prompt before it reaches the model: - **Prompt Policy** (`prompt_policy`) - enforces that the prompt adheres to a policy in line with most LLM vendors' terms (e.g. harmful content, impersonation, attempts to override the system prompt). - **PII Policy** (`pii_policy`) - enforces that the prompt contains no personally identifiable information (e.g. email addresses, phone numbers, financial details). Each entry has an `id`, human-readable `name`, `description`, and a `type` of either `input` or `output` describing which side of the model call it's applied to. Use these guardrail IDs in a task's guardrail configuration to automatically validate output. The list is the same for every caller. `org_id` and `project_id` only address the route; they do not scope the result.