Claw-code

Published on
April 11, 2026
Claw Code Source Analysis (1): Why It Is Not Just a Chat CLI but an Agent Runtime
claw-code source-code rust agents architecture
Explains why Claw Code is closer to a coding agent runtime than a simple chat CLI by looking at repository boundaries, Rust workspace design, the Python reference layer, and parity validation.
Published on
April 10, 2026
Claw Code Source Analysis (2): From the CLI Entry to the Runtime Loop
claw-code source-code rust agent-runtime cli
Breaks down how a command travels through parsing, LiveCli assembly, session/config/permission/sandbox setup, and ConversationRuntime to complete a full agent turn.
Published on
April 9, 2026
Claw Code Source Analysis (3): Tools, Skills, Slash Commands, and the Extension Systems
claw-code source-code agents tools architecture
Explains how Claw Code separates tools, skills, and slash commands into three layers and extends them into platform capabilities like plugins, MCP, workers, cron jobs, streaming, and telemetry.
Published on
April 7, 2026
Claw Code Source Analysis (4): How the Agent Manages Context and Compresses History
claw-code source-code context-management agents prompting
Explains how Claw Code manages long context through session transcripts, project context, prompt budgets, prompt cache, and session compaction instead of vector-memory retrieval.