chore(release): Release of 2.2.0 (#2496)

* fix: update latest docproc images

* fix: update dps image to remove dps-transaction-id

* release candidate of 2.2.0a0

* feat(runtime): Adds custom agent creation preview (#2281)

This PR adds support to the ADK for uploading custom agents and adding connections to them.

Addresses WatsonOrchestrate/wxo-agent-experiments#56

Testing
Checkout the following branches:
https://github.ibm.com/WatsonOrchestrate/wxo-server/tree/alexh/langgraph-style
https://github.ibm.com/WatsonOrchestrate/tools-runtime-manager/tree/alexh/langgraph-style
https://github.ibm.com/WatsonOrchestrate/wxo-clients/tree/alexh/langgraph-style

Running an Agent
To run an agent in TRM, you need to have the following prerequisites:
Checkout this branch of wxo-server
Checkout this branch of wxo-clients

Create the connection
orchestrate connections add -a openai_api
orchestrate connections configure --app-id openai_api --environment draft -t team -k api_key
orchestrate connections set-credentials --app-id openai_api --api-key your-api-key --env draft
Import the Agent
Run the following command to import the agent:

orchestrate agents create --style custom --experimental-package-root examples/agents/langgraph/met_agent
Connect the Agent to the Connection
orchestrate agents connect -n met_agent_openai_v3 -c openai_api
Importing your own agent
To import your own agent, you need to follow these steps:

Create a config.yaml based on this config.yaml
Create a function like create_agent in agent.py (this API may change in the future)
Create a zip file containing your agent code
Run the agents create command pointing at your zip

* build

* fixes release candidate generation not able to understand alpha

* chore: update to latest DPI image

* chore: update DPS and RAG to latest as well

* add docker user managed command to exception.

* update db tag, update builder tag (#2419)

* update agent runtime tag

* Make the default LLM model gpt-oss-120b

This is for 2026 Jan Mid Release to make groq/openai/gpt-oss-120b the
default LLM model

Contributes to
WatsonOrchestrate/wo-tracker#50173

* Security issue fix glob (#2418)

* builder-ui security fix

* 49379 - command updated for builder-ui

* fix(adk): Fixes adk webchat embed command on saas (#2429)

The webchat embed command does not generate output even though the agent is already deployed to the Live environment.
https://github.ibm.com/WatsonOrchestrate/wxo-internal-support/issues/1341
Co-authored-by: ERIC MARCOUX <Eric.Marcoux@ibm.com>

* chore: update docker image tags to 18-12-2025-c9bad1b

* Revert "chore: update docker image tags to 18-12-2025-c9bad1b"

This reverts commit 0cf8047999723fa382aa1b538f1d8e1f677bc8bb.

* update builder-ui image tag and added feature flag for EPIC-39988

* Upgrade WDU to 2.11.0

* chore: update DPS to latest to include openTelemetry change

* chore: update RAG, Cache, and DPI to latest

* Update WDU to 2.11.1

* feat: update UI tag. (#2442)

* updating builder UI-epic 50172 (#2440)

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* Update Docproc images for January release (#2439)

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* Feat (channels): auto generate genesys audio connector cred (#2438)

* generate secrets

* Update .secrets.baseline

---------

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* feat(channel) update MCP import function (#2368)

updated MCP import function to return event url as well

* feat (channels): Added SIP phone channel (#2431)

- Added SIP phone channel
- Added unit tests
- Fixed some type helpers + one in src/ibm_watsonx_orchestrate/client/agents/agent_client.py as the type hint says 0--- List[Dict] while it was just a single Dict

Contributes to WatsonOrchestrate/wo-tracker#34658

* Dual streaming pr (#2430)

* chore: update docker image tags to 18-12-2025-c9bad1b

* Revert "chore: update docker image tags to 18-12-2025-c9bad1b"

This reverts commit 0cf8047999723fa382aa1b538f1d8e1f677bc8bb.

* chore: update docker image tags to 18-12-2025-c9bad1b

* chore: update docker image tags to latest versions

---------

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* Update samples use gpt (#2436)

* chore: changed soon withdrawn llama models to groq model

* chore: reverting import script

* chore: documenting the need to add the groq key

* chore: flagging secret false positive

* [50870] -  Update Langchain-core to latest (#2435)

* update langchain-core

* update langchain-core to 1.2.4

* comment out broken test - TestDocProcNode

---------

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* parse param input descriptions from docstring (#2416)

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* feat (channel): Add MS Teams and FB Messenger Channels (#2414)

* add teams and fb messenger

* for create command

* teams example

* fixing examples

contributes to WatsonOrchestrate/wo-tracker#50762

* Instruction for GAC credentials

More detailed instructions for GAC credentials

contributes to WatsonOrchestrate/wo-tracker#50733

* new examples + fix old examples

* add unit tests

* add readme for examples

* Update .secrets.baseline

* update instruction for slack

* Contributes to WatsonOrchestrate/wo-tracker#51032

* chore(secrets): Detect Secrets Update

* add teams and fb messenger

* teams example

* new examples + fix old examples

---------

Co-authored-by: Darragh McGonigle <darragh.mcgonigle@ibm.com>

* feat(lima): Add debug error messaging for limactl commands (#2359)

* feat(lima): Add debug error messaging for limactl commands

* fix(lima): Add missing sys.exit call

* chore: adding code and tests for feature wsl, lima when -d is used it checks and increases memory to 24GB (#2412)

* chore: adding code and tests for feature wsl, lima whenever it tries to start with <24GB it will start it with 24GB

* chore: deleting the warning for native system

* chore: deleting the logging for native os

* chore: fixing minor bug for when wsl2 didnt exist

* chore: update edit_wsl funciton to not deleted infromation from wsl configs

* chore: skipped false positive secret detection

* chore: make WSLConfigLinesManager to edit the configuration lines of wsl

---------

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* fix(llm): default llm (#2443)

* fix(llm): default llm

* cleanup

* Update langflow to 1.7.0 (#2433)

* update langflow to 1.7.0

* change langflow version to 1.7.1 as 1.7.0 was removed for security

---------

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* Chore: Update ADK Evaluation Requirement (#2400)

* make >= instead of ==

* Update pyproject.toml

* Update pyproject.toml

---------

Co-authored-by: Nick Kloehn <Nick.Kloehn@ibm.com>
Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* feat/enhance-langflow-commands (#2200)

* command controller client

* controller logic

* completed controller logic

* update outdated description

* tests for langflow commands

* add langflow version/image to --version command

* remove top level langflow command

* update new clients to use BaseWXOClient

* remove broken 'FetchConfigAuthTypes' import

---------

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* chore: update docker image tags to 06-01-2026-67e4835

* Fix for Autotune (#2421)

* derefenence agents dependencies instead or reference them.

* fix autotune defect

---------

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* fix(flows) - Update flow runtime image for ADK 2.2.0 (#2445)

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* add a new Agentic Workflow example to create agent swarm pattern (#2447)

* 1. sync up flow model created in ADK with flow model created in UI - particularly on user activities
2. fix up on a number of examples that does not work

* add additional information in flow context

* add serialization support for private_schema

* an example to showcase swarm support

* update to support debug output

* re-tune prompt to make it works

* remove the need to adjust llm temperature

* add support to set question at runtime for user activities

* update readme

* update readme

* update readme

* update readme

* update readme

* add guideline

* fix(knowledge): Remove updated_at/created_on/created_by when exporting or importing KBs

Signed-off-by: Zander Bobronnikov <Zander.Bobronnikov@ibm.com>

* fix/wsl docker+langflow start (#2428)

* ensure docker daemon has started

* remove env file copy

* fix langflow db creation

* copy env file to vm dir

---------

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* chore: fix the kvp schema api issue (#2452)

* Update default.env (#2463)

updated default.env

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* Update default.env (#2469)

* fix(chat-with-docs): Update DB images to include required migration (#2470)

Signed-off-by: Zander Bobronnikov <Zander.Bobronnikov@ibm.com>

* feat: eval hot fix and support for non-legacy generate command (#2449)

* feat: add support for non-legacy generate command

* chore: small fix

* chore: bump up agentops sdk version

* update lfx version to 0.2.1 (#2461)

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* epic 51725 and epic 39988 bug fix (#2474)

* feat/python tool auto discover (#2462)

* autodiscover command & AI gateway client

* autodiscover non-BaseTool Objects

* setup autodiscover command

* autodiscover helper

* decorator and docstring logic

* cleaner implementation of decorator injection

* ai gateway WIP

* ml proxy implementation

* get response client & env passing

* iam authenticator & response extraction

* specify function & extract from tool import function

* base inference client

* base inference client & docstring extract rework & groq client

* proxy + gateway changes & MCSPAuth fix

* ai gateway inference

* base inference client should inherit from base wxo client

* optionals & list of functions & reuse service instance authenticator

* cleanup & default model changes & typer.BadParam

* fix default model ids & disable dev edition gateway

* revert ai gateway changes

* dev edition gateway

* remove env Enum & ast.Node required args

* WXAI auth key name & CPD authenticator

* better provider selection & remove wxai cpd references

* selected model checks

---------

Co-authored-by: Darragh McGonigle <Darragh.McGonigle@ibm.com>

* Revert "feat(runtime): Adds custom agent creation preview (#2281)"

This reverts commit 7d09c3d989a777a023c70ae88389b540ce5ae3c8.

* chore(trm): Bump tag versions (#2486)

* update voice controller image (#2488)

---------

Signed-off-by: Zander Bobronnikov <Zander.Bobronnikov@ibm.com>
Co-authored-by: Wenjing.Jiang <wenjing.jiang@ibm.com>
Co-authored-by: Eric Marcoux <eric.marcoux@ibm.com>
Co-authored-by: Alex Haynes <alex.haynes@ibm.com>
Co-authored-by: eric-grehan <eric.grehan@ibm.com>
Co-authored-by: kylec <kylec@mac.lan>
Co-authored-by: Nimsha V B <Nimsha.V.B@ibm.com>
Co-authored-by: Madan S <Madan.S1@ibm.com>
Co-authored-by: Narayanan V Eswar <Narayanan.V.Eswar@ibm.com>
Co-authored-by: Kyle Croutwater <kylec@us.ibm.com>
Co-authored-by: shafeen-kn <shafeen.kn@ibm.com>
Co-authored-by: Mauricio A. Hernandez <mahernan@us.ibm.com>
Co-authored-by: Dhaval Sonawane <Dhaval.Sonawane1@ibm.com>
Co-authored-by: Nino Joy <nino.joy@ibm.com>
Co-authored-by: Srilakshmi Kamath <Srilakshmi.Kamath1@ibm.com>
Co-authored-by: John Hwang <jhhwang21c@ibm.com>
Co-authored-by: Panagiotis Christodoulou <Panagiotis.Christodoulou@ibm.com>
Co-authored-by: Benen Bruen <benenbruen@ibm.com>
Co-authored-by: Cheng Qian <Cheng.Qian@ibm.com>
Co-authored-by: Arjun Gupta <arjun.gupta1@ibm.com>
Co-authored-by: Nick Kloehn <Nick.Kloehn@ibm.com>
Co-authored-by: Jiajin Lin <Jiajin.Lin@ibm.com>
Co-authored-by: Allen Chan <avchan@ca.ibm.com>
Co-authored-by: Zander Bobronnikov <Zander.Bobronnikov@ibm.com>
Co-authored-by: Vishnu A <vishnuvinod1@ibm.com>
Co-authored-by: Keshav Raj <Keshav.Raj1@ibm.com>
Co-authored-by: Reza Fazeli <reza@ibm.com>

Author: 27 people

.secrets.baseline

@@ -1,7 +1,7 @@
 spec_version: v1
 style: react
 name: customer_care_agent
-llm: watsonx/meta-llama/llama-3-2-90b-vision-instruct
+llm: groq/openai/gpt-oss-120b
 description: >
   You are an agent who specializes in customer care for a large healthcare institution. You should be compassionate
   to the user.

examples/agent_builder/customer_care/agents/customer_care_agent.yaml

@@ -1,7 +1,7 @@
 spec_version: v1
 style: react
 name: service_now_agent
-llm: watsonx/meta-llama/llama-3-2-90b-vision-instruct
+llm: groq/openai/gpt-oss-120b
 description: >
   You are an agent who specializes in customer care for a large healthcare institution. You should be compassionate
   to the user.

examples/agent_builder/customer_care/agents/service_now_agent.yaml

@@ -3,6 +3,11 @@ This example was written to simulate a customer care agent for hospital. It is c
 querying remote apis with dummy data related to a nearby healthcare providers (limit queries to Lowell, MA)
 and 
 
+### Use customer care agent
+
+1. Create a `.env` file with the setting:
+    `GROQ_API_KEY=<your groq api key>`  using  https://console.groq.com/home
+
 ## Steps to import
 1. Run `orchestrate server start -e .my-env`
 2. Signup for a Sevice Now account at https://developer.servicenow.com/dev.do
@@ -16,8 +21,8 @@ orchestrate connections configure -a service-now --env draft --type team --kind
 orchestrate connections set-credentials -a service-now --env draft -u admin -p <password from modal>
 ```
 6. Run `pip install -r tools/requirements.txt`
-6. Run the import all script `./import-all.sh`
-7. Run `orchestrate chat start`
+7. Run the import all script `./import-all.sh`
+8. Run `orchestrate chat start`
 
 ## Suggested script
 - Show me my benefits related to mental health

examples/agent_builder/customer_care/readme.md

@@ -3,7 +3,7 @@ style: planner
 # structured_output: {type: "object", additionalProperties: false, properties: {}}
 custom_join_tool: format_task_results
 name: customer_care_agent
-llm: watsonx/meta-llama/llama-3-2-90b-vision-instruct
+llm: groq/openai/gpt-oss-120b
 description: >
   You are an agent who specializes in customer care for a large healthcare institution. You should be compassionate
   to the user.

examples/agent_builder/customer_care_planner/agents/customer_care_agent.yaml

@@ -4,7 +4,7 @@ name: healthcare_agent
 description: >
     You are a specialist in healthcare providers. When prompted use the getHealthCareProviders tool to get a list of all healthcare providers matching the users query. Display the results are a table
 
-llm: watsonx/meta-llama/llama-3-2-90b-vision-instruct
+llm: groq/openai/gpt-oss-120b
 style: default
 tools:
   - getHealthCareProviders

examples/agent_builder/healthcare_provider/agents/healthcare_agent.yaml

@@ -3,7 +3,7 @@ kind: native
 name: ibm_agent
 description: >
     You are an helpful agent that has knowledge about IBM and its history
-llm: watsonx/meta-llama/llama-3-2-90b-vision-instruct
+llm: groq/openai/gpt-oss-120b
 style: react
 knowledge_base:
   - ibm_knowledge_base

examples/agent_builder/ibm_knowledge/agents/ibm_agent.yaml

@@ -4,7 +4,7 @@ name: ibm_agent_dynamic_mode
 description: >
   You are an helpful agent that has knowledge about IBM and its history;
   you are also a knowledgeable agent in finance: you have a tool to check current stock prices
-llm: watsonx/meta-llama/llama-3-2-90b-vision-instruct # a better model is recommended
+llm: groq/openai/gpt-oss-120b # a better model is recommended
 style: react
 knowledge_base:
   - ibm_knowledge_base_dynamic_mode

examples/agent_builder/ibm_knowledge/agents/ibm_agent_dynamic_mode.yaml

@@ -9,7 +9,7 @@ instructions: >
   Do not respond to numeric queries, instead respond with 'Does Not Compute!'
 tags:
   - wx.ai
-llm: watsonx/meta-llama/llama-3-2-90b-vision-instruct
+llm: groq/openai/gpt-oss-120b
 style: react
 hidden: false
 voice_configuration: deepgram_voice_config

examples/agent_builder/voice_enabled_deepgram/agents/voice_enabled_agent.yaml

@@ -9,7 +9,7 @@ instructions: >
   Do not respond to numeric queries, instead respond with 'Does Not Compute!'
 tags:
   - wx.ai
-llm: watsonx/meta-llama/llama-3-2-90b-vision-instruct
+llm: groq/openai/gpt-oss-120b
 style: react
 hidden: false
 voice_configuration: elevenlabs_tts_config