chore: release main

[github-actions]

🤖 I have created a release beep boop

launchdarkly-server-sdk-ai: 0.19.0

0.19.0 (2026-04-29)

⚠ BREAKING CHANGES

• Swap track_metrics_of parameter order to match spec (#144)

Features

• Add judge evaluation support to agent graphs (#142) (3d5a6a9)
• Add root-level tools map with customParameters to AI Config types (#141) (f17c535)

Bug Fixes

• Add runtime DeprecationWarnings to deprecated methods (#145) (2189b81)
• Fall back to model.parameters.tools when root tools absent (#146) (2c30d75)
• Replace done_callback with coroutine chain for judge tracking (#147) (1e1f36b)
• Swap track_metrics_of parameter order to match spec (#144) (53db736)

launchdarkly-server-sdk-ai-langchain: 0.6.0

0.6.0 (2026-04-29)

Features

• Add judge evaluation support to agent graphs (#142) (3d5a6a9)

launchdarkly-server-sdk-ai-openai: 0.5.0

0.5.0 (2026-04-29)

Features

• Add judge evaluation support to agent graphs (#142) (3d5a6a9)

---

This PR was generated with Release Please. See documentation.

---

Note

Medium Risk
Mostly a release/versioning PR, but it ships a documented breaking API change (track_metrics_of parameter order) that can impact downstream callers.

Overview
Updates release metadata and package versions for a new publish: launchdarkly-server-sdk-ai to 0.19.0, launchdarkly-server-sdk-ai-langchain to 0.6.0, and launchdarkly-server-sdk-ai-openai to 0.5.0 (manifest, pyproject.toml, and ldai.__version__).

Changelogs are regenerated to document the release contents, including a breaking change (swapped track_metrics_of parameter order) plus new judge-evaluation support for agent graphs and a few tracking/tooling-related fixes; PROVENANCE.md is updated with the new verification version.

^{Reviewed by Cursor Bugbot for commit 3ccf680. Bugbot is set up for automated code reviews on this repo. Configure here.}