release(deepagents-code): 0.1.46

[github-actions]

Caution

Merging this PR will automatically publish to PyPI and create a GitHub release.

For the full release process, see .github/RELEASING.md.

---

Release notes preview: keep this section in sync with the package CHANGELOG.md. The published GitHub release body is extracted from the merged CHANGELOG.md by release.yml, not from this PR description.

---

0.1.46 (2026-07-24)

Highlights

• Auto mode is now generally available. #4957
• Added configurable Auto goal-criteria acceptance. #4940
• Improved Auto behavior by authorizing actions from active goal/rubric directives, avoiding redundant approval prompts, showing the enable notice only on first global enable, deduplicating classifier-unavailable transcript spam, logging underlying classifier failures, and reporting classifier timeout budgets. #5017 #4993 #5012 #5013 #5011 #5025
• Added Hooks v2 capability snapshots and session transcripts, and hardened Hooks v2 command execution. #4916 #4918 #4917
• Raised the agent recursion limit to 2000 and made it configurable. #4994

Improvements and fixes

• Let the rubric grader inspect working-directory files, show rubric grader defaults, and improved /rubric help and empty-state messaging. #4835 #4966 #5015
• Unified goal activation signaling. #4980
• Made Version, Model, and CWD copyable in the Debug Console. #4975
• Improved config get output when a key is missing. #4976
• Aborted YOLO launch on Ctrl+C/Ctrl+D and made the YOLO warning friendlier for new users. #4953 #4950
• Updated LangSmith handling: secret redaction is disabled by default, /trace now flags empty env overrides that shadow the LangSmith key, and the default US endpoint is no longer treated as a custom target. #4970 #4996 #5022
• Injected OpenAI prompt_cache_key for any OpenAI-provider endpoint. #4995
• Improved tool and schema presentation: finished calls stay on the live tool-group line, first-party tool schemas now include field descriptions, and web_search/fetch_url tool descriptions were trimmed. #4927 #5019 #5016
• Omitted plugins/ and conversation_history/ from the /agent picker. #4991
• Made selector modal backdrop dimming consistent. #4990
• Restored the "Server log preserved at:" notice on exit. #4999
• Used the SDK pin as the effective editable version. #4949

End release notes preview.

---

Note

A New Contributors section is appended to the GitHub release notes automatically at publish time (see Release Pipeline, step 2).

[github-actions]

Warning

Prerelease SDK pin — Code currently pins deepagents==0.7.0b2, which is a prerelease.

This is allowed and does not block the release workflow as long as the pin is not older than the workspace SDK. Please verify this prerelease pin is intentional before merging.

Tip: if this pin is part of an intentional cross-package release sequence and the release dependency check fails because the SDK prerelease is not on PyPI yet, add the release-deps: acknowledged label to acknowledge and skip that check.

[langchain-oss-automated-triage]

Review and edit the release notes between the content markers below. Keep the version heading intact.

0.1.46 (2026-07-24)

Highlights

• Auto mode is now generally available. #4957
• Added configurable Auto goal-criteria acceptance. #4940
• Improved Auto behavior by authorizing actions from active goal/rubric directives, avoiding redundant approval prompts, showing the enable notice only on first global enable, deduplicating classifier-unavailable transcript spam, logging underlying classifier failures, and reporting classifier timeout budgets. #5017 #4993 #5012 #5013 #5011 #5025
• Added Hooks v2 capability snapshots and session transcripts, and hardened Hooks v2 command execution. #4916 #4918 #4917
• Raised the agent recursion limit to 2000 and made it configurable. #4994

Improvements and fixes

• Let the rubric grader inspect working-directory files, show rubric grader defaults, and improved /rubric help and empty-state messaging. #4835 #4966 #5015
• Unified goal activation signaling. #4980
• Made Version, Model, and CWD copyable in the Debug Console. #4975
• Improved config get output when a key is missing. #4976
• Aborted YOLO launch on Ctrl+C/Ctrl+D and made the YOLO warning friendlier for new users. #4953 #4950
• Updated LangSmith handling: secret redaction is disabled by default, /trace now flags empty env overrides that shadow the LangSmith key, and the default US endpoint is no longer treated as a custom target. #4970 #4996 #5022
• Injected OpenAI prompt_cache_key for any OpenAI-provider endpoint. #4995
• Improved tool and schema presentation: finished calls stay on the live tool-group line, first-party tool schemas now include field descriptions, and web_search/fetch_url tool descriptions were trimmed. #4927 #5019 #5016
• Omitted plugins/ and conversation_history/ from the /agent picker. #4991
• Made selector modal backdrop dimming consistent. #4990
• Restored the "Server log preserved at:" notice on exit. #4999
• Used the SDK pin as the effective editable version. #4949

When the release changes are finalized, run @dcode-release-bot apply. Merge only after the curated release-notes check passes.

If new relevant entries appear after applying, run @dcode-release-bot draft and then @dcode-release-bot apply again.

The only bypass is the release: dangerously skip curated notes label.

[open-swe]

✅ Open SWE Review: No issues found

Open SWE reviewed this PR and found no potential bugs to report.

Open in Web • View Open SWE trace

[Mason Daugherty (mdrxy)]

@dcode-release-bot apply

[langchain-oss-automated-triage]

Curated release notes were applied to the package changelog and release PR body.

Do not add more deepagents-code changes before merge unless you are prepared to run @dcode-release-bot draft and @dcode-release-bot apply again.

[open-swe]

Open SWE Review found 1 potential issue.

Open in Web • View Open SWE trace

[github-actions]

The package release workflow has started:

• release(deepagents-code): 0.1.46

Follow the linked run for build, test, and publish status.