v1.10.3

web-infra-dev/midscenev1.10.3Jul 7, 2026by github-actions[bot]

AI Summary

Focuses on reporting and CLI improvements, offering better visibility into LLM usage and export capabilities while cleaning up internal parsing logic.

Key Highlights

Added agent markdown export view for reports.
Exposed aggregated LLM usage metrics on the agent.
Simplified model JSON parsing logic.
Added model check command to the CLI.

New Features

Markdown export view
Aggregated LLM usage metrics
Model check command

Full Release Notes

<!-- Release notes generated using configuration in .github/release.yml at v1.10.3 -->

## What's Changed
### Other Changes
* feat(report): add agent markdown export view by @quanru in https://github.com/web-infra-dev/midscene/pull/2711
* refactor(core): simplify model JSON parsing by @EAGzzyCSL in https://github.com/web-infra-dev/midscene/pull/2690
* feat(core): expose aggregated LLM usage metrics on agent (#1084) by @quanru in https://github.com/web-infra-dev/midscene/pull/2740
* fix(core): handle malformed Doubao bbox responses by @EAGzzyCSL in https://github.com/web-infra-dev/midscene/pull/2758
* docs(shared): update common tools comment by @EAGzzyCSL in https://github.com/web-infra-dev/midscene/pull/2734
* fix(core): preserve report response dump shape by @EAGzzyCSL in https://github.com/web-infra-dev/midscene/pull/2760
* feat(cli): add model check command  by @EAGzzyCSL in https://github.com/web-infra-dev/midscene/pull/2722


**Full Changelog**: https://github.com/web-infra-dev/midscene/compare/v1.10.2...v1.10.3