🌀
← Back to Articles & Artefacts
artefactseast

CLAUDE

IAIP Research
agent-orchestration

CLAUDE

Inquiry github issue: miadisabelle/Etuaptmumk-RSM#145 source github issue: jgwill/src#372 (the PDE Evals....)

Motivation

  • An article received that talks about "Composio Open Sources Agent Orchestrator to Help AI Developers Build Scalable Multi-Agent Workflows Beyond the Traditional ReAct Loops"

Sources and files

  • from https://www.perplexity.ai/search/agent-orchestration-b611db20-a-IDD.OpW6TnC0rQKkhydwcg
    • agent-orchestration--b611db20-a2c2-42ef-84cc-a060b.md the produce file to help develop

What would this help advance ??

/src/Miadi/ and make it capable to orchestrate our work

/src/Miadi/COORDINATION_MerryPoppingParasol.md

  • An envisionned plan that would create a containerized and more efficient ways to work within Miadi.
    • Might invalidate or need next point bellow about A2A...

/src/Miadi/A2A_*

  • Was an incompleted approach for making this orchestration possible
    • Either upgrade it or edit it if that becames invalid or not a great path

/workspace/repos/jgwill/medicine-wheel

  • Might need to consider why Langchain/LangGraph in relation to what was created in /workspace/repos/jgwill/medicine-wheel so we can install what is needed in /src/Miadi/ to orchestrate...

/workspace/repos/avadisabelle/lang*js

  • Any of our ongoing langgraph/langchain libraries to help either Miadi or other platform have great packages todo the work

AVA.md

  • She is helping me on claude.ai to write some narrative about it... and ./output/260224132923.txt contains what I sent to that thread to continue working,
    • That could really be part of what Mia ingest for working on something in the stack... ../MWDS-3e7eef71-7fc9-40aa-b24e-f0b37ef71caf--2602211121--closing/ might help get somewhere because it closed some development cycle in a new ways that would relate to this.

mnemoniclab-story-stcmastery-langgraph-miaco-260224121247.md is the synopsis I gave to Ava

  • See the APPENDIX section of AVA.md so LLM should know what to work on...

compass_artifact_wf-84ca4f83-183e-4e46-b705-44b2dbe9004a_text_markdown.md

This document was produced by Ava as that is her message:

Your AI cognitive support systems research report is ready. It covers four core technical patterns—multi-format RAG, multi-agent orchestration, human-in-the-loop gates, and adaptive routing—with production case studies from Unstructured.io, Abridge, LinkedIn, Viz.ai, Not Diamond, Khanmigo, and others. Key findings include why the ingestion layer (not the vector store) is the true RAG differentiator, why production multi-agent systems rely on deterministic backbones over autonomous agents, and why user-state-aware adaptive orchestration remains the least mature and most novel engineering challenge in the space.

/src/pde-evals/260224--50a54b2f-88ac-4edb-899c-5c7f9edb5fe5/

  • A potential location where sessions were created for exploring the role of the "Prompt-Decomposition-Engine" in what we are doing... that would affect and help this develop