deepagents-code==0.1.46

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

---

Internal maintainers: Nick Hollon (@nick-hollon-lc), Johannes du Plessis (@johannes117), Alexander Olsen (@aolsenjazz), ccurme (@ccurme), Shrikar Seshadri (@davibinboi), Mason Daugherty (@mdrxy)

Released by: Mason Daugherty (@mdrxy)

Git log since deepagents-code==0.1.45

This commit history includes changes to this package. Commits are listed newest first.

• efa86c5 release(deepagents-code): 0.1.46 (#4965)
• e1b1303 chore(code): bump SDK pin to 0.7.0b2 and dependency minimums (#5031)
• 516619b fix(code): report Auto classifier timeout budget (#5025)
• 5dbedcd docs(code): document dev vs GA tracing projects (#4973)
• 170b4a6 fix(code): trim web_search and fetch_url tool descriptions (#5016)
• b4d39a4 refactor(code): drop redundant goal-tool system prompt injection (#5023)
• f6caf52 refactor(code): remove get_rubric source metadata (#5024)
• ffb7aab fix(code): ignore LangSmith default US endpoint as custom target (#5022)
• e63cddf style(code): tighten Auto mode first-run notice copy (#5021)
• 4ccb873 fix(code): put field descriptions on first-party tool schemas (#5019)
• c2dbd17 fix(code): show Auto enable notice modal only on first global enable (#5012)
• 9500306 fix(code): dedupe Auto classifier-unavailable transcript spam (#5013)
• 728dbe7 fix(code): authorize Auto actions from active goal/rubric directives (#5017)
• 1a3f46b fix(code): improve /rubric help and empty-state messaging (#5015)
• 6d6632f fix(code): inject OpenAI prompt_cache_key for any OpenAI-provider endpoint (#4995)
• bd03e8a fix(code): log underlying Auto classifier failures (#5011)
• 1709dbf fix(code): make Auto mode generally available (#4957)
• 313f952 fix(code): flag empty env override shadowing LangSmith key in /trace (#4996)
• 976c54b fix(code): keep finished calls on live tool-group line (#4927)
• 52b8993 feat(code): make Version, Model, CWD copyable in Debug Console (#4975)
• 260c33e fix(code): give config get a useful missing-key message (#4976)
• f792279 docs(code): document first-token shell allow-list bypass (T13) (#4942)
• 5bc7714 fix(code): restore "Server log preserved at:" notice on exit (#4999)
• 37addfc feat(code): configure Auto goal criteria acceptance (#4940)
• ca86fbe feat(code): add Hooks v2 session transcripts (#4918)
• 9af90ae chore(code): bump SDK pin to 0.7.0b1 and dependency minimums (#5000)
• 94962a5 feat(code): raise agent recursion limit to 2000 and make it configurable (#4994)
• 3606e3f fix(code): consistent modal backdrop dimming for selectors (#4990)
• db26688 fix(code): omit plugins/ and conversation_history/ from /agent picker (#4991)
• 9340518 feat(sdk,code,quickjs)!: make the ToDoListMiddleware list opt-in (#4929)
• 9975874 fix(code): avoid redundant Auto approval prompts (#4993)
• 0c4016e fix(code,evals): unify goal activation signaling (#4980)
• 05dd0d1 fix(code): harden Hooks v2 command execution (#4917)
• 5dcf749 feat(code): let the rubric grader inspect working-directory files (#4835)
• 465b9de fix(code): show rubric grader defaults (#4966)
• d046427 revert(code,sdk): revert SystemPromptConfig (#4969)
• 8c31545 style(code): render -m slash prompts as plain agent text (#4974)
• 1b270c8 fix(code): use SDK pin as effective editable version (#4949)
• 8b4e508 feat(code): add Hooks v2 capability snapshots (#4916)
• e6ab9c0 fix(code): abort YOLO launch on Ctrl+C/Ctrl+D (#4953)
• d029585 fix(code): disable LangSmith secret redaction by default (#4970)
• fe07f39 fix(code): make YOLO warning friendlier for new users (#4950)
• 28d138a style(code): hide terminal cursor in pre-TUI selectors (#4951)
• a8d1b32 feat(sdk,code,quickjs): lean system prompt by default, restorable (#4859)
• 803a27f refactor(code): use concrete middleware backends (#4963)