梦兽编程™AI_SUITE [ SECTION / AI ]
Claude Code
Claude Code tips, internals, and practical guides.
48 POSTS01
AI / 2026.08.15 / 8 MIN
Harness Engineering Principles - Architecture Wisdom from Claude Code
Distilling Claude Code's six core Harness Engineering principles: prompts as control plane, cache-aware design, fail-closed, A/B testing everything, observe before …
↗
02
AI / 2026.08.08 / 9 MIN
Cross-Session Memory - From Forgetting to Persistent Learning
Deep dive into Claude Code's cross-session memory system: six-layer memory architecture, automatic memory extraction, session summarization, and nightly consolidation …
↗
03
AI / 2026.08.01 / 6 MIN
89 Feature Flags - Claude Code's 'Secret Switches'
Deep dive into Claude Code's Feature Flag system: the design philosophy behind 89 functional switches, five functional domain categorizations, and how build-time switches …
↗
04
AI / 2026.07.25 / 6 MIN
Effort, Fast Mode, and Thinking - AI's 'Thinking Depth' Control Knob
Deep dive into Claude Code's reasoning control mechanisms: Effort levels, Fast Mode acceleration, and Thinking mode - plus how to adjust AI thinking depth based on task …
↗
05
AI / 2026.07.18 / 7 MIN
Agent Clusters - When AI Starts 'Forming Squads'
Deep dive into Claude Code's multi-Agent architecture: three spawning methods - sub-Agents, Fork mode, and Coordinator mode - plus Agent Swarm collaboration and …
↗
06
AI / 2026.07.11 / 8 MIN
Microcompact - The Art of Precise Context Pruning
A deep dive into Claude Code's Microcompact mechanism: differences from auto-compaction, precise pruning strategies, relevance filtering, and fine-grained context …
↗
07
AI / 2026.07.11 / 5 MIN
Hooks - Jumping In at Key Points
A deep dive into Claude Code's Hooks system: 26 lifecycle events, four hook types, execution models and exit code semantics, and how to customize AI Agent behavior.
↗
08
AI / 2026.07.04 / 6 MIN
YOLO Classifier - AI Deciding 'Can I Do This?'
A deep dive into Claude Code's YOLO classifier mechanism: how AI autonomously judges operation safety, two-stage review process, safety whitelist and rejection tracking, …
↗
09
AI / 2026.06.20 / 5 MIN
Skills System——Teaching AI Your "Secret Techniques"
Deep dive into Claude Code's skills system: built-in vs user-defined skills, how SkillTool works, differences and cooperation with MCP tools. Unveiling how to teach AI …
↗
10
AI / 2026.06.20 / 7 MIN
Cache Optimization Patterns - Making Cache Hit Rates Soar
A deep dive into Claude Code's cache optimization strategies: tool ordering stability, MCP tool management, session preservation techniques, and maximizing prompt cache …
↗
AI / 2026.08.15 / 8 MIN
Harness Engineering Principles - Architecture Wisdom from Claude Code
Distilling Claude Code's six core Harness Engineering principles: prompts as control plane, cache-aware design, fail-closed, A/B testing everything, observe before …
↗
02
AI / 2026.08.08 / 9 MIN
Cross-Session Memory - From Forgetting to Persistent Learning
Deep dive into Claude Code's cross-session memory system: six-layer memory architecture, automatic memory extraction, session summarization, and nightly consolidation …
↗
03
AI / 2026.08.01 / 6 MIN
89 Feature Flags - Claude Code's 'Secret Switches'
Deep dive into Claude Code's Feature Flag system: the design philosophy behind 89 functional switches, five functional domain categorizations, and how build-time switches …
↗
04
AI / 2026.07.25 / 6 MIN
Effort, Fast Mode, and Thinking - AI's 'Thinking Depth' Control Knob
Deep dive into Claude Code's reasoning control mechanisms: Effort levels, Fast Mode acceleration, and Thinking mode - plus how to adjust AI thinking depth based on task …
↗
05
AI / 2026.07.18 / 7 MIN
Agent Clusters - When AI Starts 'Forming Squads'
Deep dive into Claude Code's multi-Agent architecture: three spawning methods - sub-Agents, Fork mode, and Coordinator mode - plus Agent Swarm collaboration and …
↗
06
AI / 2026.07.11 / 8 MIN
Microcompact - The Art of Precise Context Pruning
A deep dive into Claude Code's Microcompact mechanism: differences from auto-compaction, precise pruning strategies, relevance filtering, and fine-grained context …
↗
07
AI / 2026.07.11 / 5 MIN
Hooks - Jumping In at Key Points
A deep dive into Claude Code's Hooks system: 26 lifecycle events, four hook types, execution models and exit code semantics, and how to customize AI Agent behavior.
↗
08
AI / 2026.07.04 / 6 MIN
YOLO Classifier - AI Deciding 'Can I Do This?'
A deep dive into Claude Code's YOLO classifier mechanism: how AI autonomously judges operation safety, two-stage review process, safety whitelist and rejection tracking, …
↗
09
AI / 2026.06.20 / 5 MIN
Skills System——Teaching AI Your "Secret Techniques"
Deep dive into Claude Code's skills system: built-in vs user-defined skills, how SkillTool works, differences and cooperation with MCP tools. Unveiling how to teach AI …
↗
10
AI / 2026.06.20 / 7 MIN
Cache Optimization Patterns - Making Cache Hit Rates Soar
A deep dive into Claude Code's cache optimization strategies: tool ordering stability, MCP tool management, session preservation techniques, and maximizing prompt cache …
↗