neuro-san==0.6.7 Now with full neuro-san message support in tracing/Observability apps

[d1donlydfink]

Apps like LangSmith, Arize Phoenix, Honey Hive now have full neuro-san message support. Basically everything you get in the agent_thinking files from agent_cli can now appear in the tracing, like:

• AGENT_FRAMEWORK messages that span the entire agent network request
• AGENT_TOOL_RESULT messages for neuro-san level interpretation of tool results
• AGENT messages for various other tool calling at the neuro-san level.
• AGENT_PROGRESS messages from coded tools

Note that while sly_data will show up in the initial and end AGENT_FRAMEWORK messages, the keys present are visible but the values are redacted.