Releases: cognizant-ai-lab/neuro-san
neuro-san 0.6.24
What's Changed
- Add documentation showing a method to connect to MCP server with authentication by @Noravee in #651
- NS-641 Remove colons from filename to fix hocon test cases on Windows by @d1donlydfink in #653
- temporarily disabled test case by @vince-leaf in #656
- NS-615 Extend CLI client to support MCP sessions. by @andreidenissov-cog in #652
- Add "runtime" to the policy object keys to remove from messages by @Noravee in #659
Full Changelog: 0.6.23...0.6.24
neuro-san 0.6.23
What's Changed
- Extend MCP tool call parameters. by @andreidenissov-cog in #588
- Add gemini 3 flash model in default_llm_info.hocon by @Noravee in #645
- Add thinking parameters for Gemini models by @Noravee in #646
- Update copyrights by @d1donlydfink in #649
- Allow multiple manifest files in AGENT_MANIFEST_FILE env var by @d1donlydfink in #648
Note:
There is a small logging change that shifts the log message telling you when the next read of the manifest files will be.
This used to be at INFO level and now it is at DEBUG level. This increases the signal-to-noise for the vast majority of users
in default logs where the periodicity for checking manifest files is ~1 second by making it disappear by default.
If you cared about this message, please modify the loggers section of your logging.json for your deployment accordingly with something like:
"neuro_san.service.watcher.registries.registry_storage_updater.RegistryStorageUpdater": {
"level": "DEBUG",
"handlers": [
"http_console"
],
"propagate": false
},
See https://docs.python.org/3/library/logging.config.html#dictionary-schema-details
as to how this file can be configured for your own needs. Examples there are provided in YAML,
but these can be easily translated to JSON (which we prefer).
Another good resource: https://docs.python.org/3/howto/logging-cookbook.html
Full Changelog: 0.6.22...0.6.23
neuro-san 0.6.22
What's Changed
- Update dependencies cap to langchain-core -> 2.0 for dependabot fix by @d1donlydfink in #643
Full Changelog: 0.6.21...0.6.22
neuro-san 0.6.21
What's Changed
- Upgrade langchain-openai to a minimum of 1.0.0 to match langchain and langchain-core by @ofrancon in #637
- NS-638: Fix conversion for ChatMessages. by @andreidenissov-cog in #639
Full Changelog: 0.6.20...0.6.21
neuro-san 0.6.20
What's Changed
- Allow agents called from CodedTools to be discoverable by validation by @d1donlydfink in #636
Full Changelog: 0.6.19...0.6.20
neuro-san 0.6.19
What's Changed
- Fix async issues : delete_resources() not awaited and AsyncDirectAgentSession not getting answers by @d1donlydfink in #631
Full Changelog: 0.6.18...0.6.19
neuro-san 0.6.18
What's Changed
- Update to leaf-server-common==0.1.23 by @d1donlydfink in #629
Full Changelog: 0.6.17...0.6.18
neuro-san 0.6.17
What's Changed
- Some build tweaks by @d1donlydfink in #624
- Add gpt 5.2 and gemini 3 pro by @Noravee in #625
Full Changelog: 0.6.16...0.6.17
neuro-san 0.6.16
What's Changed
- NS-584 Update BaseMessageDictionaryConverter to allow AgentProgressMessages to come through by @d1donlydfink in #618
AGENT_PROGRESS messages now show up in the agent_thinking files for agent_cli (among other things).
Full Changelog: 0.6.15...0.6.16
neuro-san 0.6.15
What's Changed
- NS-472 Use latest langchain-nvidia-ai-endpoints which is compatible with langchain v1 by @d1donlydfink in #596
- Add Claude Opus 4.5 in default_llm_info.hocon by @Noravee in #599
- MCP Docs - First version. by @andreidenissov-cog in #598
- Tweak agent hocon reference docs to emphasize essentials for beginners by @d1donlydfink in #617
Full Changelog: 0.6.14...0.6.15