feat: Add MLflow Prompt Registry provider

This commit adds a new remote provider for the Prompts API that integrates
with MLflow's Prompt Registry (MLflow 3.4+), enabling centralized prompt
management and versioning.

Key features:
- Full CRUD operations (create, read, update, list)
- Immutable version control with default version management
- Automatic variable extraction from {{ variable }} placeholders
- Deterministic ID mapping between Llama Stack and MLflow
- Comprehensive test coverage (41 unit + 14 integration tests)

Provider implementation:
- src/llama_stack/providers/remote/prompts/mlflow/
  - config.py: Configuration schema with validation
  - mapping.py: ID mapping and metadata utilities
  - mlflow.py: Main adapter implementation (520 lines)
- src/llama_stack/providers/registry/prompts.py: Provider registration

Testing:
- tests/unit/providers/remote/prompts/mlflow/
  - test_config.py: 14 configuration tests
  - test_mapping.py: 27 ID mapping tests
- tests/integration/providers/remote/prompts/mlflow/
  - test_end_to_end.py: 14 end-to-end scenarios
  - conftest.py: Test fixtures and server availability checks
  - README.md: Testing documentation
- scripts/test_mlflow_prompts_manual.py: Manual validation script

Documentation:
- docs/docs/providers/prompts/remote_mlflow.mdx: User guide
- docs/docs/providers/prompts/index.mdx: Provider category index

Dependencies:
- Requires mlflow>=3.4.0 (added to provider pip_packages)

Limitations:
- No deletion support (MLflow constraint)
- Sequential version numbers only (1, 2, 3...)

Signed-off-by: William Caban <[email protected]>
Co-Authored-By: Claude <[email protected]>

feat: Add inline::reference provider and MLflow authentication

This commit addresses the PR feedback to refactor prompts architecture
and add authentication support for the MLflow provider.

- Created new inline::reference provider in providers/inline/prompts/reference/
- Moved implementation from core/prompts/prompts.py to reference provider
- Updated core/prompts/prompts.py to delegate to inline provider for backward compatibility
- Follows Llama Stack pattern where core routes, providers implement

- Registered inline::reference provider (InlineProviderSpec)
   - Zero dependencies, uses KVStore backend (SQLite, PostgreSQL, etc.)
   - Full CRUD support including deletion
- Registered remote::mlflow provider (RemoteProviderSpec)
   - Requires mlflow>=3.4.0
   - Added provider_data_validator for authentication support

Added three-tier authentication support following inference provider pattern:

1. **Per-request provider data** (highest priority)
   - Via x-llamastack-provider-data header
   - Enables multi-tenant deployments with user-specific credentials

2. **Configuration auth_credential** (fallback)
   - Server-level authentication in config file
   - Supports environment variable substitution

3. **Environment variables** (lowest priority)
   - Standard MLflow behavior (MLFLOW_TRACKING_TOKEN, etc.)

- Added MLflowProviderDataValidator for request header validation
- Added auth_credential field (SecretStr) to MLflowPromptsConfig
- MLflowPromptsAdapter inherits from NeedsRequestProviderData
- Implemented _get_mlflow_token() method with proper precedence
- Added @json_schema_type decorator and sample_run_config() method

- Complete documentation for inline::reference provider
- Configuration examples for SQLite and PostgreSQL
- Migration guide from core implementation
- Best practices and troubleshooting

- Added comprehensive Authentication section
- Documented all three authentication methods with examples
- Added Databricks authentication guide
- Added enterprise MLflow examples
- Security best practices
- Updated configuration table and sample configs

- Added "Available Providers" section
- Added "Choosing a Provider" decision guide
- Quick start examples for both providers
- Common features overview

- Updated test_config.py with authentication tests
   - MLflowProviderDataValidator tests
   - auth_credential configuration tests
   - sample_run_config() tests

- inline::reference unit tests: 12/12
- remote::mlflow unit tests: 49/49
- remote::mlflow integration tests: 14/14

All existing imports continue to work:
- from llama_stack.core.prompts import PromptServiceConfig
- from llama_stack.core.prompts import PromptServiceImpl
- from llama_stack.core.prompts import get_provider_impl

Prompts work without MLflow installed (uses inline::reference)

Signed-off-by: William Caban <[email protected]>
Co-Authored-By: Claude <[email protected]>

Author: williamcaban

docs/docs/providers/prompts/index.mdx

@@ -0,0 +1,93 @@
+---
+sidebar_label: Prompts
+title: Prompts
+---
+
+# Prompts
+
+## Overview
+
+This section contains documentation for all available providers for the **prompts** API.
+
+The Prompts API enables centralized management of prompt templates with versioning, variable handling, and team collaboration capabilities.
+
+## Available Providers
+
+### Inline Providers
+
+Inline providers run in the same process as the Llama Stack server and require no external dependencies:
+
+- **[inline::reference](inline_reference.mdx)** - Reference implementation using KVStore backend (SQLite, PostgreSQL, etc.)
+  - Zero external dependencies
+  - Supports local SQLite or PostgreSQL storage
+  - Full CRUD operations including deletion
+  - Ideal for local development and single-server deployments
+
+### Remote Providers
+
+Remote providers connect to external services for centralized prompt management:
+
+- **[remote::mlflow](remote_mlflow.mdx)** - MLflow Prompt Registry integration (requires MLflow 3.4+)
+  - Centralized prompt management across teams
+  - Built-in versioning and audit trail
+  - Supports authentication (per-request, config, or environment variables)
+  - Integrates with Databricks and enterprise MLflow deployments
+  - Ideal for team collaboration and production environments
+
+## Choosing a Provider
+
+### Use `inline::reference` when:
+- Developing locally or deploying to a single server
+- You want zero external dependencies
+- SQLite or PostgreSQL storage is sufficient
+- You need full CRUD operations (including deletion)
+- You prefer simple configuration
+
+### Use `remote::mlflow` when:
+- Working in a team environment with multiple users
+- You need centralized prompt management
+- Integration with existing MLflow infrastructure
+- You want to leverage Databricks or enterprise MLflow
+- You need authentication and multi-tenant support
+- Advanced versioning and audit trail capabilities are required
+
+## Quick Start Examples
+
+### Using inline::reference
+
+```yaml
+prompts:
+  - provider_id: local-prompts
+    provider_type: inline::reference
+    config:
+      run_config:
+        storage:
+          stores:
+            prompts:
+              type: sqlite
+              db_path: ./prompts.db
+```
+
+### Using remote::mlflow
+
+```yaml
+prompts:
+  - provider_id: mlflow-prompts
+    provider_type: remote::mlflow
+    config:
+      mlflow_tracking_uri: http://localhost:5555
+      experiment_name: llama-stack-prompts
+      auth_credential: ${env.MLFLOW_TRACKING_TOKEN}
+```
+
+## Common Features
+
+All prompt providers support:
+- Create and store prompts with version control
+- Retrieve prompts by ID and version
+- Update prompts (creates new versions)
+- List all prompts or versions of a specific prompt
+- Set default version for a prompt
+- Automatic variable extraction from `{{ variable }}` templates
+
+For detailed documentation on each provider, see the individual provider pages linked above.