BLOG

Knowledge Share

Technical articles, tutorials, and insights

Found 345 posts · Page 1/29
Code AgentHarness EngineeringContext Engineering

Code Agent Anatomy (12): Harness Design Part 2 — Context Engineering

A harness engineering perspective on MyCodeAgent's context engineering: why History and ModelView are separated, read-time projection vs. write-time deletion, dual-source token estimation for compaction decisions, and proactive + reactive trigger paths. This is the second post in Part 4 Harness Engineering, focused on 'how the agent decides what to show the model.'

·9 min read
Code AgentHarness EngineeringTool Pipeline

Code Agent Anatomy (13): Harness Design Part 3 — Tool Execution Pipeline

A harness engineering perspective on MyCodeAgent's tool execution pipeline: concurrent grouping with ordering guarantees, the four-gate execution pipeline (permission → optimistic lock → circuit breaker → execute), and two-layer byte budget control. This is the third post in Part 4 Harness Engineering, focused on 'what happens between a model tool_call output and the result being written to history.'

·8 min read
open-sourcelinuxdesktop

Open Source Project #198: Omarchy — DHH's Opinionated Linux Desktop, Arch + Hyprland, 30k Stars

A fully curated Linux desktop distribution by DHH (creator of Ruby on Rails), built on Arch Linux + Hyprland tiling WM + Quickshell. Ships with 22 beautiful themes, Neovim, first-class AI coding agent support, a complete dev environment, and a keyboard-first operating philosophy. Shell, MIT, 30k Stars.

·10 min read
open-sourcedocumentationagent

Open Source Project #199: OpenWiki — LangChain's Self-Maintaining Codebase Wiki CLI, Built for Agents, 15k Stars

LangChain AI's open-source CLI for automatic codebase documentation generation and maintenance. An agent reads your source code and writes a structured Markdown wiki, tracks each material fact back to exact code lines (Grounded Claims), and automatically updates when the code changes. Supports 13 model providers, Claude Code/Codex/OpenCode integrations, 9 knowledge connectors, and an interactive node-graph visualizer. TypeScript, MIT, 15k Stars.

·9 min read
open-sourcedigital-humantalking-head

Open Source Project #196: LiveTalking — Real-Time Interactive Streaming Digital Human Engine, Wav2Lip/MuseTalk/ER-NeRF, 9k Stars

Real-time interactive streaming digital human engine widely deployed in production. Supports Wav2Lip, MuseTalk, ER-NeRF, and Ultralight-Digital-Human rendering models. Features: voice cloning, interrupt-and-resume dialog, full-body video compositing, WebRTC/RTMP/virtual camera output, and multi-session concurrency. Core pipeline: text/voice input → LLM response → TTS synthesis → lip-sync inference → audio/video stream. Plugin architecture for TTS/Avatar/Output modules. Python, Apache-2.0, 9k Stars.

·9 min read
open-sourceagent-skillsevaluation

Open Source Project #197: skill-up — Alibaba's Agent Skills Evaluation and Evolution Tool, Closed-Loop Testing with Auto-Repair

Alibaba's open-source evaluation and evolution tool for Agent Skills. Evaluation: declarative YAML config (eval.yaml + cases/*.yaml), multi-engine support (Claude Code/Codex/Qoder CLI/Qwen Code), three judging strategies (rule/script/agent judge), Anthropic-compatible reports. Evolution: built-in skill-upper Agent Skill reads failure reports, auto-repairs the Skill or expands eval cases, reruns iteratively until passing. GitHub Action CI support. Go, Apache-2.0, 655 Stars.

·10 min read
Code AgentHarness EngineeringControl Flow

Code Agent Dissection (11): Harness Design Part 1 — Control Flow

Examining MyCodeAgent's control flow from a harness engineering design perspective: why a single main loop rather than multiple nested loops, the engineering value of immutable state machines, the completion gate's feedback loop design, and the completeness of termination paths. This is the first article of Part 4 Harness Engineering, focused on 'why it was designed this way' rather than 'what it is.'

·8 min read
Code AgentContext EngineeringToken Budget

Code Agent Dissection (09): When the Conversation Gets Too Long, What Happens When Tokens Run Out?

A deep dive into MyCodeAgent's context engineering: HistoryManager's append-only fact log, ProjectionBuilder's read-time projection, ContextBudgetPolicy's trigger logic, and ContextCompactor's LLM summary compression. Understand why 'history is never deleted' and 'giving the model a bounded view' are two completely different things, and how an agent compresses context within budget without losing facts.

·8 min read
Code AgentTranscriptPersistence

Code Agent Dissection (10): When the Agent Crashes, How Does It Recover? Where Is Conversation History Stored?

A deep dive into MyCodeAgent's persistence and recovery mechanism: TranscriptStore's append-only JSONL event stream, five event types, ResumeLoader rebuilding runtime state from events, UncertainAction's tool interruption handling, and SessionMemory for cross-run feed-forward memory. Understand why agents can't rely on memory for history, and the complete chain for deterministic crash recovery.

·9 min read
enterprise knowledge baseGraphRAGHippoRAG

Enterprise Knowledge Base (03): Graph-Enhanced RAG Benchmark — GraphRAG vs HippoRAG

Same 89 questions, two graph-enhanced RAG frameworks: GraphRAG 3.1.1 and HippoRAG 2.0. This article covers the full deployment journey — GLM-4-flash structured output failures, LanceDB dimension mismatches, NV-Embed-v2 offline loading — and the actual numbers across single-hop, multi-hop, and boundary refusal.

·9 min read
Code AgentSubagentTask Tool

Code Agent Dissection (08): When a Task Is Too Complex, How Do You Delegate to a Sub-Agent?

A deep dive into MyCodeAgent's sub-agent mechanism: the Task tool, RuntimeProfile sandbox constraints, SubagentLauncher's complete execution chain, and structured result contracts. Understand how the main agent delegates exploratory tasks to a read-only lightweight sub-agent, and why sub-agents are not 'simplified main agents' but strictly constrained execution units.

·13 min read
Code AgentMCPTool Adapter

Code Agent Dissection (07): How Are External Tools Integrated? A Deep Dive into MCP

A deep dive into MyCodeAgent's MCP integration: optional dependencies, config discovery, stdio/HTTP transport, Adapter disguising remote tools as local Tools, and result normalization into a unified protocol. Understand how external capabilities enter the same tool pipeline, and why 'registered in the registry' and 'allowed to execute' are two separate concerns.

·11 min read