claude-code-source-analysis (xiaonancs)
Summary
A systems-engineering reconstruction of the complete Claude Code agent harness from a reverse-engineered v2.1.88 snapshot (includes the extracted source zip plus appendix analyses). Part of a series that also covers OpenAI Codex CLI internals, giving a comparative harness view.
Key Takeaways
- Reconstructs the full harness from an actual extracted v2.1.88 snapshot
- Sister studies (Codex CLI) enable cross-harness comparison
- Strong on context management and compaction mechanics
Reliability Note
Independent reverse-engineering effort - unofficial, verify against official Anthropic docs and behavior.
Flows informed by this source
4
Implement Context Compaction
Keep long agent sessions alive by compacting old conversation into structured summaries before the context window fills.
+1 more steps to Done
Best forproduction-grade builds with strict verification
Session Summaries That Survive Compaction
Write durable session summaries at the right moments so knowledge outlives compaction and process restarts.
+1 more steps to Done
Best forbuilders who have shipped a basic app before
Prompt-Cache-Aware Conversation Design
Structure your agent's context for prefix caching: stable prefixes, append-only history, and cache-friendly dynamic content.
Best forbuilders who have shipped a basic app before
Tool Result Size Management
Stop tool outputs from flooding context: budgets per tool, smart truncation, pagination, and reference-based results.
+1 more steps to Done
Best forbuilders who have shipped a basic app before