Shipped · Observability

What mlflow/mlflow shipped

The public repository of Mlflow · github.com/mlflow/mlflow

Written by FoxPlug from the repository's own releases; not affiliated with Mlflow. Paste your own repository in the box below and FoxPlug writes the same weekly update for your product. This page is an automatic summary of the releases, pull requests and commits in github.com/mlflow/mlflow; Mlflow did not write it and does not use or endorse FoxPlug. Every line links to the change it describes.

Get a weekly update like this for your product, free

Or Use it as a GitHub Action

See the update FoxPlug writes for mlflow, free

Week of November 2, 2026

What shipped

Why it matters

This week adds automatic issue detection for agents in the 3.18.0 release and tightens security around model code execution and permission grants. It also fixes several AI Gateway problems, including endpoint setup for SAP AI Core and errors after secret deletion. Together these changes make agent failures easier to find and gateway endpoints more reliable to configure.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

MLflow 3.18.0 is out with automatic issue detection that groups recurring agent failures in an Issues tab. Also this week: model code execution checks before dependency install, SAP AI Core endpoints in the AI Gateway, and fixes for gateway secrets and provider configs.

Post for LinkedIn

MLflow 3.18.0 is released. It adds automatic issue detection, which turns recurring agent failures into issues you can investigate and fix from a dedicated Issues tab. This week also brings security and correctness work: PyFunc environment preparation now enforces MLFLOW_ALLOW_MODEL_CODE_EXECUTION before installing model-supplied dependencies, and model or prompt deletion is rejected while a named role holds an exact-name allow grant. On the AI Gateway side, you can now create an SAP AI Core endpoint from the UI, endpoint API key errors after secret deletion are fixed, and provider region and endpoint settings are validated.

Week of October 26, 2026

What shipped

Why it matters

This week fixed a file path traversal issue in FileStore and stopped the client from resending POST writes that may already have been committed. Spark UDF environment restoration on Spark Connect and CrewAI autologging on 1.15.29 were both repaired, and the store backends now agree on trace id validation and assessment handling.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

This week: FileStore path traversal fix, client stops resending POST writes after timeouts, spark_udf archives work with Spark Connect, CrewAI 1.15.29 autologging restored, and a new issue detection run history page. Plus a Rename action in the evaluation runs table.

Post for LinkedIn

Changes this week in MLflow: FileStore now rejects trace and artifact paths built from ids that could escape the store root, addressing CVE-2026-53995. MlflowClient stopped resending POST writes that the server may have already processed after a 5xx response or read timeout. spark_udf environment archives no longer carry absolute bin/python symlinks, so Spark Connect safe extraction can restore virtualenv and uv environments. CrewAI autologging works again with CrewAI 1.15.29. A new issue detection run history page lists detection runs, failed runs and runs with no findings. SqlAlchemyStore now validates trace ids the same way FileStore does, FileStore.start_trace stores submitted assessments, and system trace tags are handled consistently when traces start. The evaluation runs table gaine

Week of October 19, 2026

What shipped

Changelog entry

Week of October 12, 2026

What shipped

Why it matters

This week adds the skill registry and lets custom scorers in job containers run for both online trace and session scoring and automatic evaluation, extending how evaluation can be applied. It also records dataset, scorer and agent lineage on evaluation runs so users can see what an evaluation covered. Alongside that, a batch of bug fixes addresses prompt formatting, trace ordering, duplicate span export and Spark study and Optuna trial state issues that affected day-to-day use.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

MLflow 3.17.1 is out with minor features, fixes and docs updates. Also this week: the skill registry lands, custom scorers can run in job containers for online trace and session scoring and automatic evaluation, evaluation runs record dataset/scorer/agent lineage, and fixes cover

Post for LinkedIn

This week MLflow shipped 3.17.1, a patch release with minor feature enhancements, bug fixes and documentation updates, including Decision API support in make_judge(). The skill registry was added following RFC-0008. Custom scorers can now run in job containers for online trace scoring, online session scoring and automatic evaluation. Evaluation runs record lineage for the dataset, scorers and agent, session_id can be a list for hierarchical sessions, and the evaluation runs page table, toolbar and chart were refreshed behind a feature flag. Fixes this week cover format_prompt rewriting placeholders inside variable values, search_traces ordering by end_time in the SQL store, duplicate span export in strands, semantic_kernel and haystack autolog with a shared tracer provider, and Spark study

Week of October 5, 2026

What shipped

Changelog entry

Week of September 28, 2026

What shipped

Why it matters

This week adds significant TypeScript SDK capabilities with TypeSafe AI integration and autologging, improves documentation for custom tracing features, and fixes multiple streaming issues in tool call handling across providers. Several artifact and endpoint fixes address data handling and UTF-8 encoding problems.

Changelog entry

Week of September 21, 2026

What shipped

Why it matters

This week brings infrastructure improvements for trace analytics and job execution extensibility, alongside fixes for common issues in model registration, tracing across multiple LLM providers, and data handling in pandas DataFrames. The SQL optimizations and plugin architecture lay groundwork for scaling production deployments.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

MLflow 2.11.5: opt-in UC model registry artifact routing via Databricks SDK Files API, SQL trace analytics optimizations, job executor plugins, and Skill Registry artifact lifecycle support.

Week of September 14, 2026

What shipped

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

MLflow 3.16.1 is out with bug fixes, security updates, and improvements to gateway model discovery and UC trace handling.

Weeks with too little public activity are left out rather than filled in. Last updated 2026-11-09.

Is this your repo? Ask us to remove this page.