deepagents-talon==0.0.7

### Highlights

- Added Discord channel support, including improved gateway failure reporting after startup. ([#5992](https://github.com/langchain-ai/deepagents/issues/5992), [#6113](https://github.com/langchain-ai/deepagents/issues/6113))
- Added chat-scoped conversation history with configurable storage and optional hybrid search. ([#6105](https://github.com/langchain-ai/deepagents/issues/6105), [#6115](https://github.com/langchain-ai/deepagents/issues/6115), [#6108](https://github.com/langchain-ai/deepagents/issues/6108))
- Added MCP configuration tools, channel-based MCP server authorization, hot reload for MCP configuration, and OAuth/device authentication support for Slack and GitHub MCP integrations. ([#6097](https://github.com/langchain-ai/deepagents/issues/6097), [#6073](https://github.com/langchain-ai/deepagents/issues/6073), [#6084](https://github.com/langchain-ai/deepagents/issues/6084), [#6078](https://github.com/langchain-ai/deepagents/issues/6078), [#6079](https://github.com/langchain-ai/deepagents/issues/6079))
- Added more flexible subagent execution, including background expendable subagents, dcode-style subagents in fork mode, fresh subagents, per-task tool selection, and on-demand subagent configuration reloads. ([#6098](https://github.com/langchain-ai/deepagents/issues/6098), [#6085](https://github.com/langchain-ai/deepagents/issues/6085), [#6129](https://github.com/langchain-ai/deepagents/issues/6129), [#6099](https://github.com/langchain-ai/deepagents/issues/6099))
- Added defensive research defaults and a configuration-hardening self-review skill, with missing research subagent defaults backfilled. ([#6131](https://github.com/langchain-ai/deepagents/issues/6131), [#6136](https://github.com/langchain-ai/deepagents/issues/6136), [#6135](https://github.com/langchain-ai/deepagents/issues/6135))

### New features

- Added a `/help` command. ([#6106](https://github.com/langchain-ai/deepagents/issues/6106))
- Added a `current_time` tool and timezone-aware wall-clock cron schedules. ([#6065](https://github.com/langchain-ai/deepagents/issues/6065), [#6062](https://github.com/langchain-ai/deepagents/issues/6062))
- Added persistent LangGraph checkpoints. ([#6088](https://github.com/langchain-ai/deepagents/issues/6088))
- Added channel debug logging and opt-in agent activity logging. ([#5983](https://github.com/langchain-ai/deepagents/issues/5983), [#5984](https://github.com/langchain-ai/deepagents/issues/5984))
- New messages can now interrupt active turns. ([#6023](https://github.com/langchain-ai/deepagents/issues/6023))
- Long agent turns now keep the typing indicator alive. ([#5993](https://github.com/langchain-ai/deepagents/issues/5993))

### Fixes and improvements

- Improved channel reconnect resilience. ([#6040](https://github.com/langchain-ai/deepagents/issues/6040))
- Improved cron reliability by keeping the ticker alive after failed ticks, and stored cron jobs in a structured, versioned format. ([#6087](https://github.com/langchain-ai/deepagents/issues/6087), [#6086](https://github.com/langchain-ai/deepagents/issues/6086))
- Improved conversation history embeddings and search by correcting token budgets and prompt defaults, making remote embedding settings explicit, keeping search non-blocking, and removing the default Qwen query prefix. ([#6132](https://github.com/langchain-ai/deepagents/issues/6132), [#6133](https://github.com/langchain-ai/deepagents/issues/6133), [#6134](https://github.com/langchain-ai/deepagents/issues/6134))
- Fixed OAuth and MCP integration issues, including TLS hostname normalization, omitted empty optional MCP arguments, persisted OAuth token expiry, secured OAuth discovery, and restarted token refresh. ([#6102](https://github.com/langchain-ai/deepagents/issues/6102), [#6077](https://github.com/langchain-ai/deepagents/issues/6077), [#6090](https://github.com/langchain-ai/deepagents/issues/6090), [#6100](https://github.com/langchain-ai/deepagents/issues/6100))
- Fixed WhatsApp behavior by restoring bridge compatibility, preserving quoted message context and approval loops, handling reactions, and restricting replies to self-chat. ([#5999](https://github.com/langchain-ai/deepagents/issues/5999), [#6025](https://github.com/langchain-ai/deepagents/issues/6025), [#6104](https://github.com/langchain-ai/deepagents/issues/6104), [#6010](https://github.com/langchain-ai/deepagents/issues/6010))
- Treated trailing `[SILENT]` as a suppression marker. ([#6110](https://github.com/langchain-ai/deepagents/issues/6110))

---

Internal maintainers: @jkennedyvz, @linted

Released by: @jkennedyvz

<details>
<summary>Git log since deepagents-talon==0.0.6</summary>

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

- [`80898ad`](https://github.com/langchain-ai/deepagents/commit/80898ade19fd42171d3222bfeaa7d945fe68fe85) release(deepagents-talon): 0.0.7 (#5989)
- [`e760662`](https://github.com/langchain-ai/deepagents/commit/e760662c956c581a0aca2cd4c95a3a7ce2cbd187) feat(talon): add configuration hardening self-review skill (#6136)
- [`54cf354`](https://github.com/langchain-ai/deepagents/commit/54cf354d04d094eaf77c17ed239fdd235e4ebc83) fix(talon): backfill missing research subagent defaults (#6135)
- [`1a835d2`](https://github.com/langchain-ai/deepagents/commit/1a835d216260d3b140af528d42fbb75eca26f818) feat(talon): ship defensive research defaults (#6131)
- [`3698f10`](https://github.com/langchain-ai/deepagents/commit/3698f10bc46ef6ac6d7b30c1fcdc499a8ac08b28) fix(talon): make remote embedding settings explicit and search non-blocking (#6133)
- [`276a385`](https://github.com/langchain-ai/deepagents/commit/276a385d6c49ef63f0cae107e5a1afa591191e03) feat(talon): support fresh subagents and per-task tool selection (#6129)
- [`933d863`](https://github.com/langchain-ai/deepagents/commit/933d86350f0d0ffa54359d4f00510e52ccdc2b85) fix(talon): stop HistoryEmbeddings defaulting to Qwen's query prefix (#6134)
- [`caabca2`](https://github.com/langchain-ai/deepagents/commit/caabca2b479ec731a81fa42e4c53c9471895fd7b) fix(talon): correct history embedding token budget and prompt defaults (#6132)
- [`904326a`](https://github.com/langchain-ai/deepagents/commit/904326a7932f72a189fcf0a74ba1f77a4087f79f) feat(talon): add optional hybrid conversation history search (#6108)
- [`99fcef8`](https://github.com/langchain-ai/deepagents/commit/99fcef85410736e211f5b0bcb2e9f1651e685105) feat(talon): add configurable conversation history storage (#6115)
- [`07d2952`](https://github.com/langchain-ai/deepagents/commit/07d2952d346d81d06bd181db8c560a77f2b51bc8) chore(deps): raise dependency minimums for `all` (#6096)
- [`94e4520`](https://github.com/langchain-ai/deepagents/commit/94e452077cf2650b0a410da7fa1e23b28d5dd8e1) fix(talon): report Discord gateway failures after startup (#6113)
- [`92576da`](https://github.com/langchain-ai/deepagents/commit/92576daf1cfaadfc811445b29be9457a78eb4e5e) fix(talon): treat trailing [SILENT] as a suppression marker (#6110)
- [`6c89fe2`](https://github.com/langchain-ai/deepagents/commit/6c89fe2197a2dfe4f3851cda38565bcadba6066b) release(langchain-quickjs): 0.3.7 (#6095)
- [`0b7b3a5`](https://github.com/langchain-ai/deepagents/commit/0b7b3a5b7b49e7ef9c4cb6775ba4e2d6a9b84f43) refactor(talon): extract conversation command dispatch (#6107)
- [`d121141`](https://github.com/langchain-ai/deepagents/commit/d12114143ec4ec15a04c5a7970c82899bf1b3caa) feat(talon): add chat-scoped conversation history (#6105)
- [`919ad2b`](https://github.com/langchain-ai/deepagents/commit/919ad2b56335c4e4cc416d474745efd78e0324a3) feat(talon): add /help command (#6106)
- [`eca6203`](https://github.com/langchain-ai/deepagents/commit/eca6203338a4d7927456ff1ec70b3a29f01e9799) fix(talon): preserve WhatsApp approval loops and handle reactions (#6104)
- [`34005a4`](https://github.com/langchain-ai/deepagents/commit/34005a4efab0ab2ee2466fec413844ddae0b27c1) fix(talon): normalize OAuth TLS server hostname (#6102)
- [`bf78360`](https://github.com/langchain-ai/deepagents/commit/bf78360c87725a3af546e97d03915fb64ecf13c0) feat(talon): reload subagent configuration on demand (#6099)
- [`c7a4841`](https://github.com/langchain-ai/deepagents/commit/c7a48414e90456bc6a453a3cbb60cc787221fc74) feat(talon): run expendable subagents in the background (#6098)
- [`43c2994`](https://github.com/langchain-ai/deepagents/commit/43c299460042b8ae44b2162fba24be456088bdb9) fix(talon): secure OAuth discovery and restart token refresh (#6100)
- [`f5f8dfe`](https://github.com/langchain-ai/deepagents/commit/f5f8dfec051905befc83f65ff40100fd01ad2865) feat(talon): add MCP configuration tools with default approval (#6097)
- [`91bcf63`](https://github.com/langchain-ai/deepagents/commit/91bcf63805c656fe0399aeaf9d25c0ae7f7a2cf2) feat(talon): support dcode-style subagents in fork mode (#6085)
- [`337416a`](https://github.com/langchain-ai/deepagents/commit/337416a85800818a5660afe8cfaedca16f8ac085) feat(talon): persist LangGraph checkpoints (#6088)
- [`1ac5386`](https://github.com/langchain-ai/deepagents/commit/1ac5386790ad2a069923161393410ef49a39a2b0) feat(talon): reload MCP configuration without restarting (#6084)
- [`a02d2df`](https://github.com/langchain-ai/deepagents/commit/a02d2df874332784530082261f801512bbf62dee) fix(talon): persist OAuth token expiry (#6090)
- [`4e5f935`](https://github.com/langchain-ai/deepagents/commit/4e5f9350e4d77b8bf19e472e8414662d3fa59dc0) perf(talon): store cron jobs in a structured, versioned format (#6086)
- [`4c062ec`](https://github.com/langchain-ai/deepagents/commit/4c062ec76e2cf19abb3ac8a78b1bf68cf7ece3cd) fix(talon): keep the cron ticker alive through a failed tick (#6087)
- [`d91ec7d`](https://github.com/langchain-ai/deepagents/commit/d91ec7d0a7b8d0a703fa16b5ee4237c84d10526c) feat(talon): support GitHub MCP device authentication (#6079)
- [`f8b137e`](https://github.com/langchain-ai/deepagents/commit/f8b137ef22bf81f8178fe490f6dad7621f161a16) feat(talon): support Slack MCP OAuth login (#6078)
- [`0fe5f94`](https://github.com/langchain-ai/deepagents/commit/0fe5f942d6359f956614584f37137bbba78ab589) feat(talon): add a current_time tool (#6065)
- [`632f2c9`](https://github.com/langchain-ai/deepagents/commit/632f2c941b877eff70407606b58e393212448a26) fix(talon): omit empty optional MCP arguments (#6077)
- [`76dd223`](https://github.com/langchain-ai/deepagents/commit/76dd22334c95083449cc858dec245f98a383bba4) feat(talon): authorize MCP servers through channels (#6073)
- [`83caa95`](https://github.com/langchain-ai/deepagents/commit/83caa9579dc1be8b2336ab194caaf4b77e859730) refactor(talon): own MCP OAuth login (#6070)
- [`43fe6dd`](https://github.com/langchain-ai/deepagents/commit/43fe6dde63831ac089cf15a14cc2932a20daabd3) refactor(talon): own MCP runtime configuration (#6066)
- [`afff4f8`](https://github.com/langchain-ai/deepagents/commit/afff4f8841bcc370d01c2739f53e882a92ef5c6a) feat(talon): add timezone-aware wall-clock cron schedules (#6062)
- [`3812967`](https://github.com/langchain-ai/deepagents/commit/3812967c84d90619848f3185f22470204fc88bd8) release(deepagents-code): 0.1.66 (#5963)
- [`bdccb82`](https://github.com/langchain-ai/deepagents/commit/bdccb82d21f998174354a25daa4624dc0d9da97b) chore(deps): raise dependency minimums for `all` (#6053)
- [`ec28fc1`](https://github.com/langchain-ai/deepagents/commit/ec28fc184e1356567bb10bac24eb95cec146b95d) release(langchain-quickjs): 0.3.6 (#5607)
- [`ac0fac4`](https://github.com/langchain-ai/deepagents/commit/ac0fac42f6ece4147c3636896358e1a9391cd94a) chore(deps): raise dependency minimums for `all` (#6046)
- [`54fe91f`](https://github.com/langchain-ai/deepagents/commit/54fe91fd3745e285899961bfe74380c837674164) fix(talon): improve channel reconnect resilience (#6040)
- [`1aae374`](https://github.com/langchain-ai/deepagents/commit/1aae3746682a65c837c5dd0f165b685253fe9465) release(deepagents): 0.7.13 (#6031)
- [`3db758c`](https://github.com/langchain-ai/deepagents/commit/3db758c6672713b7ad05dc0933f83ea6298797d5) chore(deps): raise dependency minimums for `all` (#6012)
- [`03436b3`](https://github.com/langchain-ai/deepagents/commit/03436b369c0324498602fe6b7918cf36f3629d76) fix(talon): preserve WhatsApp quoted message context (#6025)
- [`f93752f`](https://github.com/langchain-ai/deepagents/commit/f93752f855836b2af358a7c71c35a9f774af58e3) feat(talon): interrupt active turns for new messages (#6023)
- [`a197807`](https://github.com/langchain-ai/deepagents/commit/a19780791e020ccb772701285c459fb6a4dab0ff) release(deepagents): 0.7.12 (#6006)
- [`e0d0afa`](https://github.com/langchain-ai/deepagents/commit/e0d0afa18a3b8209f20c24cfda2ad766e476a102) feat(talon): keep typing indicator alive during long agent turns (#5993)
- [`67c6c17`](https://github.com/langchain-ai/deepagents/commit/67c6c17a316c84a6f21eafb95f12e56eedf11e7e) feat(talon): add Discord channel adapter (#5992)
- [`40359ec`](https://github.com/langchain-ai/deepagents/commit/40359ec683eaff3a67b39d3f6b3003e70db9ec4d) fix(talon): restrict WhatsApp replies to self-chat (#6010)
- [`568b398`](https://github.com/langchain-ai/deepagents/commit/568b398df9b9f4f3464b4107c0ef9001f530d728) fix(talon): restore WhatsApp bridge compatibility (#5999)
- [`3a0f68c`](https://github.com/langchain-ai/deepagents/commit/3a0f68ccd08166394e02fd736869482be5759f83) feat(talon): add opt-in agent activity logging (#5984)
- [`1c14626`](https://github.com/langchain-ai/deepagents/commit/1c14626d068ee5e1724d53dc17927d37987ebe36) feat(talon): add channel debug logging (#5983)
- [`4bc1100`](https://github.com/langchain-ai/deepagents/commit/4bc11004ba86999c2f6d59b1d958b53195619bcc) release(deepagents): 0.7.11 (#5977)
- [`a233ded`](https://github.com/langchain-ai/deepagents/commit/a233ded7cffd4cc5c81e5767780bf93472dd8fe7) release(deepagents-code): 0.1.65 (#5929)
- [`77b0f01`](https://github.com/langchain-ai/deepagents/commit/77b0f018f9d736c9409e44d03edaf9985b59f271) chore(deps): raise dependency minimums for `all` (#5936)
- [`d8686f7`](https://github.com/langchain-ai/deepagents/commit/d8686f74df7b8f838ec6679086c6291e3096a39e) release(deepagents-code): 0.1.64 (#5848)
- [`d26b53a`](https://github.com/langchain-ai/deepagents/commit/d26b53ad12265375c6dcf6b1c1501e0a4adf0ea3) release(deepagents): 0.7.10 (#5838)

</details>