Module Dependency Graph
Info: Auto-regenerated section The Mermaid graph between the AUTO markers is rewritten by
scripts/sync_wiki.pyfrom the actual intra-project imports on every sync. Do not hand-edit inside the markers.
flowchart LR
a2a_server[a2a_server.py]
agent[agent.py]
app[app.py]
callbacks[callbacks.py]
clinical_schemas[clinical_schemas.py]
config[config.py]
context[context.py]
database[database.py]
document_processor[document_processor.py]
human_in_the_loop[human_in_the_loop.py]
llm[llm.py]
memory[memory.py]
mock_data[mock_data.py]
models[models.py]
observability[observability.py]
orchestration[orchestration.py]
plugins[plugins.py]
prompts[prompts.py]
security[security.py]
tools[tools.py]
vector_store[vector_store.py]
agent --> a2a_server
agent --> app
callbacks --> agent
callbacks --> orchestration
callbacks --> plugins
clinical_schemas --> orchestration
clinical_schemas --> tools
config --> a2a_server
config --> agent
config --> app
config --> callbacks
config --> context
config --> document_processor
config --> llm
config --> memory
config --> observability
config --> plugins
config --> security
config --> tools
config --> vector_store
context --> agent
context --> plugins
database --> document_processor
database --> mock_data
document_processor --> tools
human_in_the_loop --> agent
llm --> agent
llm --> orchestration
llm --> tools
memory --> agent
memory --> orchestration
models --> tools
observability --> agent
observability --> callbacks
observability --> human_in_the_loop
observability --> memory
observability --> plugins
observability --> tools
orchestration --> agent
plugins --> app
prompts --> agent
prompts --> orchestration
security --> callbacks
security --> memory
security --> plugins
tools --> agent
tools --> orchestration
Design property
No circular dependencies. config.py and prompts.py are standalone; everything converges on agent.py, which wires the root agent, then app.py (runtime) and a2a_server.py (A2A serving) consume it.
Related: Module Reference · Agent Architecture