ashwin-magalu/claw-code

How ashwin-magalu/claw-code works

This repository is primarily a Python porting workspace, not yet a runtime-equivalent agent harness. Its main workflow is to invoke the CLI, assemble a workspace manifest, and produce summaries or other views of mirrored commands, tools, and runtime behavior. The Python runtime models routing and turns, while query/session modules provide turn output and persistence. The README also identifies parity auditing and inventory views as workspace capabilities. Rust crates are present as an in-progress port, but the available documentation centers the active Python workspace; unsampled Python subsystems are shown only where the CLI or source index supports them.

Main entry points: main.py Python CLI · query_engine.py Query engine · runtime.py Port runtime · tools.py Tool inventory

Components 15

CLI and reporting

Python CLI src/main.py

Port inventory

Port manifest src/port_manifest.py
Port models src/models.py
Command inventory src/commands.py
Tool inventory src/tools.py
Tool permissions src/permissions.py

Mirrored runtime

Query engine src/query_engine.py
Port runtime src/runtime.py
Runtime context src/context.py

Workspace state

Turn history src/history.py
Session store src/session_store.py
Transcript store src/transcript.py
Setup report src/setup.py
Parity audit src/parity_audit.py

Other components

CLI user

Connections 21

CLI user → Python CLI: invokes README.md
Python CLI → Port manifest: builds src/main.py
Python CLI → Query engine: renders summary src/main.py
Python CLI → Port runtime: runs turns src/main.py
Python CLI → Command inventory: lists commands src/main.py
Python CLI → Tool inventory: lists tools src/main.py
Python CLI → Tool permissions: configures permissions src/main.py
Python CLI → Parity audit: runs audit src/main.py
Python CLI → Setup report: reports setup src/main.py
Query engine → Port manifest: builds workspace view src/query_engine.py
Query engine → Command inventory: loads backlog src/query_engine.py
Query engine → Tool inventory: loads backlog src/query_engine.py
Query engine → Port models: uses data types src/query_engine.py
Query engine → Session store: loads and saves src/query_engine.py
Query engine → Transcript store: records turns src/query_engine.py
Port runtime → Query engine: runs turn engine src/runtime.py
Port runtime → Command inventory: routes commands src/runtime.py
Port runtime → Tool inventory: routes tools src/runtime.py
Port runtime → Runtime context: builds context src/runtime.py
Port runtime → Turn history: records events src/runtime.py
Tool inventory → Tool permissions: filters tools src/tools.py

More Rust diagrams: 0xkoiner/deterministic-deployer-evm · aerosim-open/aerosim · aptos-labs/aptos-indexer-processor-sdk · aranya-project/aranya-core · ereinaimer/taurine

Drawn by GitDiagram on September 30, 2026 from the repository's file tree, README and source files. Read as Markdown ·