singhcoder/claude-code

Claude Code Snapshot for Research. All original source code is the property of Anthropic.

Structural Verdict

How Data Flows Through the System

Commands flow through argument parsing to query engines that route against archived reference data, producing manifest summaries and parity audits.

1. Parse CLI args — argparse processes subcommands and arguments from user input
2. Load reference data — JSON snapshots of commands and tools are loaded with caching
3. Build manifest — Scan Python workspace to create current state summary (config: src_root, total_python_files)
4. Route queries — Match user prompts against archived command/tool inventories with scoring
5. Generate output — Render markdown summaries, audit reports, or query results

System Behavior

How the system actually operates at runtime — where data accumulates, what loops, what waits, and what controls what.

Data Pools

Delays & Async Processing

• JSON File Loading (async-processing, ~milliseconds) — First access to command/tool data requires file system read and JSON parsing
• Filesystem Scanning (async-processing, ~seconds) — Manifest building scans entire Python workspace directory tree

Control Points

• Query Result Limit (runtime-toggle) — Controls: Number of results returned by commands and tools subcommands. Default: 20 (default), 32 (subsystems), 5 (route)
• Cache Size (threshold) — Controls: LRU cache size for snapshot loading functions. Default: maxsize=1
• Python Version Requirement (env-var) — Controls: Minimum Python version for the porting workspace. Default: 3.13+

Technology Stack

Key Components

• main (function) — CLI entry point that routes subcommands for summary, manifest, parity audit, and query operations src/main.py
• QueryEnginePort (class) — Core query engine that renders markdown summaries and routes prompts through the ported system src/query_engine.py
• PortRuntime (class) — Routes user prompts against mirrored command/tool inventories with scoring src/runtime.py
• build_port_manifest (function) — Scans the Python workspace and builds a manifest of all modules and their metadata src/port_manifest.py
• run_parity_audit (function) — Compares the Python workspace against the archived TypeScript source for coverage analysis src/parity_audit.py
• load_command_snapshot (function) — Loads mirrored command entries from archived JSON snapshot with LRU caching src/commands.py
• load_tool_snapshot (function) — Loads mirrored tool entries from archived JSON snapshot with LRU caching src/tools.py
• PortingModule (class) — Data model representing a module in the porting process with name, responsibility, and status src/models.py
• CostTracker (class) — Tracks resource usage units and events during porting operations src/cost_tracker.py
• build_port_context (function) — Creates a PortContext with paths to source, tests, and assets directories src/context.py

Configuration

src/context.py (python-dataclass)

• source_root (Path, unknown)
• tests_root (Path, unknown)
• assets_root (Path, unknown)

src/cost_tracker.py (python-dataclass)

• total_units (int, unknown) — default: 0
• events (list[str], unknown) — default: field(default_factory=list)

src/dialogLaunchers.py (python-dataclass)

• name (str, unknown)
• description (str, unknown)

src/history.py (python-dataclass)

• title (str, unknown)
• detail (str, unknown)

Related Ml Inference Repositories

Frequently Asked Questions

Analyzed on April 1, 2026 by CodeSea. Written by Karolina Sarna.