awesome-repositories.com
Blog
MCP
awesome-repositories.com

Discover the best open-source repositories with AI-powered search.

ExploreCurated searchesOpen-source alternativesSelf-hosted softwareBlogSitemap
ProjectMCP serverAboutHow we rankPress
LegalPrivacyTerms
© 2026 Bringes Technology SRL·VAT RO45896025·hello@awesome-repositories.com
code-yeongyu avatar

code-yeongyu/lazycodex

0
View on GitHub↗
lazycodex.ai↗

Lazycodex

Lazycodex is a plugin package that installs an agent harness into the Codex coding assistant with a single command. It adds planning, execution, review and diagnostic workflows to existing sessions, built around a plan-execute-verify loop: a decision-complete plan is drafted before any code is written, carried out step by step with progress saved after each item, and repeated until an independent reviewer confirms with concrete evidence that the work is finished.

The harness also provides multi-agent orchestration. Specialist roles such as explorer, librarian and planner are adopted by child agents, which can run in parallel, each with its own instructions and model. Tasks are routed to models by category, so the strongest reasoning model is kept for hard work and cheaper models handle small edits, with ordered fallback chains when a preferred model is unavailable.

Beyond orchestration, it gives agents persistent project memory through hierarchical guidance files written beside the code they describe. It adds language-server diagnostics, references and renames, along with structural search and rewrite across a codebase. A library of skills and pre- and post-action lifecycle hooks can be extended, and the defaults can be overridden. Review and cleanup workflows run after implementation without changing behavior.

The plugin installs through a package-runner command with no global install, and it also supports an unattended mode and installation from a Git-hosted marketplace. Hooks require explicit approval before they run. Final configuration finishes in a background worker during the first approved session, followed by a restart prompt. A health report and a clean uninstall command are included.

Features

  • Codex Plugin Installation - Installs an agent harness into Codex with a single command so existing sessions gain extra workflows without manual setup.
  • Coding Agent Plugin Packages - Packages planning, execution, review and diagnostic workflows as a plugin installed into Codex with one package-runner command.
  • Agent Iteration Loops - Repeats an agent task up to an iteration cap, stopping when a configurable completion phrase confirms the reviewer's verdict.
  • Agent Safety and Verified Completion - Keeps tasks running until an independent reviewer confirms completion with concrete evidence instead of accepting unverified claims of success.

AI search

Explore more awesome repositories

Describe what you need in plain English — the AI ranks thousands of curated open-source projects by relevance.

Start searching with AI
3,889 stars·252 forks·TypeScript·MIT·1 view
  • Role-Based Agent Orchestration - Defines specialist roles like explorer, librarian and planner that child agents adopt, each with its own instructions and model.
  • Parent Agent Sub-Agent Delegation - Defines specialist roles such as explorer, librarian and planner that child agents adopt, each with its own instructions and model.
  • Model Routing Categories - Maps each task category to an ordered model chain, using strong reasoning models for hard work and cheaper ones for small edits.
  • Multi-Agent Orchestration Frameworks - Delegates subtasks to specialist child agents such as explorer, librarian and planner, running in parallel, each with its own instructions and model.
  • Multi-Model Orchestration - Delegates subtasks to parallel specialist agents and routes each to a suitable model by category, saving the strongest model for hard work.
  • Native Harness Plugin Installs - Packages extra planning, execution and review workflows as a native plugin that the Codex coding assistant loads.
  • Step-by-Step Task Plans - Breaks large codebase changes into a decision-complete multi-step plan before any code is written.
  • Verified Agent Task Loops - Repeating an agent task in a loop and executing a saved plan until an independent reviewer confirms with concrete evidence that the work is truly finished.
  • Implementation Planning - Produces a decision-complete plan file from a task request without touching product code, settling every choice before implementation.
  • Agentic Plan-And-Execute Workflows - Drafts a decision-complete plan first, executes it step by step with saved progress, and loops until an independent reviewer confirms completion.
  • Guidance Generators - Generates hierarchical guidance files beside the directories that need them, giving agents local landmarks before they edit code.
  • Agent Lifecycle Hooks - Runs pre- and post-action hooks around agent steps to report status, enforce checks and inject behavior.
  • Agent Skill Libraries - Ships an extensible library of reusable skills and pre- and post-action lifecycle hooks whose defaults can be overridden.
  • Agentic Project Memories - Writes hierarchical guidance files beside the code they describe, giving agents local landmarks and persistent project conventions before editing.
  • Model Fallback Chains - Defines ordered fallback chains per task category so the agent switches to another available model when the preferred one cannot be used.
  • Model Selection for Delegated Tasks - Runs several agents in parallel and automatically selects a suitable model for each delegated subtask by task category.
  • Project Memory Retention - Keeps persistent project memory available in assistant sessions so agents can recall accumulated knowledge about a codebase.
  • Saved Plan Execution - Carries out a saved plan, optionally in a separate worktree, until every checkbox is done, then reports that orchestration is complete.
  • Semantic Code Analyzers - Offers language-server diagnostics, definitions, references, symbols and renames, plus structural search and rewrite across a codebase.
  • Codebase Intelligence - Gives agents hierarchical project memory files, language-server navigation and structural search and rewrite for exploring and understanding large codebases.
  • Agent Skill Extenders - Ships an extensible skill library and lifecycle hooks with overridable defaults that extend agent behavior without core changes.
  • Codebase-Driven Plan Generators - Drafts a decision-complete plan of ordered steps from the codebase before any product code is written.

    • Star history

    Star history chart for code-yeongyu/lazycodexStar history chart for code-yeongyu/lazycodex

    How this analysis was created: This summary and feature list are AI-generated from collected project material and can contain mistakes. Stars, license and language are imported from GitHub. Inclusion does not mean that we have tested or audited this project. Check the source documentation for any feature you depend on. Learn more on our About page.

    Projects sharing features with Lazycodex

    These projects share indexed features with Lazycodex. Shared tags can include platform or build tooling; verify the primary use case before treating a result as a replacement.
    • lintsinghua/claude-code-booklintsinghua avatar

      lintsinghua/claude-code-book

      4,359View on GitHub↗

      This repository is a long-form technical book about how a production AI coding agent runtime, often called an agent harness, is built. It dissects a real-world coding agent's architecture chapter by chapter and turns it into transferable lessons, and it also works as a step-by-step guide to designing and building your own harness around a language model. Python scripts in the repository support the book itself rather than the agent. The book's core subject is the machinery around the model: a streaming conversation loop that repeats tool calls until the model returns plain text, a tool protoc

      Pythonagentagent-architectureai-agent
      View on GitHub↗4,359
    • michael-denyer/pstack-claudemichael-denyer avatar

      michael-denyer/pstack-claude

      4,131View on GitHub↗

      pstack-claude is a plugin and skills pack for AI coding assistants. It bundles rigorous engineering workflows, step-by-step playbooks, routing hooks, model settings and subagent definitions, and translates them so the same content runs natively on many coding harnesses. Its skills are derived from upstream sources, with local forks and translations preserved and licensing provenance recorded. The distinctive parts are cross-harness porting and multi-agent orchestration. Tool names, model names and subagent types are pretranslated into each runtime's equivalents before installation, so nothing

      JavaScriptagent-pluginagent-skillsagentic-ai
      View on GitHub↗4,131
    • xiaomimimo/mimo-codeXiaomiMiMo avatar

      XiaomiMiMo/MiMo-Code

      13,767View on GitHub↗

      MiMo-Code is a terminal-based AI coding assistant that reads and edits project files, runs shell commands, manages version control, and delegates work to LLM-powered agents from a single interactive interface. It connects to major model providers through API keys, browser sign-in, or imported credentials, and accepts any OpenAI-compatible custom endpoint defined by a base URL and model ID. Much of what sets it apart is how it handles long-running work. Project knowledge, session checkpoints, notes, and task logs are kept in a searchable database and injected back into context when a session r

      TypeScriptaiai-agentscli
      View on GitHub↗13,767
    • claude-code-best/claude-codeclaude-code-best avatar

      claude-code-best/claude-code

      20,272View on GitHub↗

      Claude Code is a command-line interface and multi-agent orchestration framework designed for autonomous software engineering. It enables AI agents to perform codebase modifications, debugging, and Git workflow management while coordinating multiple specialized agents to decompose and execute complex engineering tasks in parallel. The system distinguishes itself through a high degree of isolation and safety, utilizing Git worktrees to create independent working directories for concurrent agents and implementing a tiered permission system that combines user rules, project policies, and OS-level

      TypeScript
      View on GitHub↗20,272
    Compare all 30 related projects→

    Frequently asked questions

    What does code-yeongyu/lazycodex do?

    Lazycodex is a plugin package that installs an agent harness into the Codex coding assistant with a single command. It adds planning, execution, review and diagnostic workflows to existing sessions, built around a plan-execute-verify loop: a decision-complete plan is drafted before any code is written, carried out step by step with progress saved after each item, and repeated until an independent reviewer confirms with concrete evidence that the work is finished.

    What are the main features of code-yeongyu/lazycodex?

    The main features of code-yeongyu/lazycodex are: Codex Plugin Installation, Coding Agent Plugin Packages, Agent Iteration Loops, Agent Safety and Verified Completion, Role-Based Agent Orchestration, Parent Agent Sub-Agent Delegation, Model Routing Categories, Multi-Agent Orchestration Frameworks.

    Which projects share features with code-yeongyu/lazycodex?

    Projects with overlapping indexed features include: lintsinghua/claude-code-book — This repository is a long-form technical book about how a production AI coding agent runtime, often called an agent… michael-denyer/pstack-claude — pstack-claude is a plugin and skills pack for AI coding assistants. It bundles rigorous engineering workflows,… xiaomimimo/mimo-code — MiMo-Code is a terminal-based AI coding assistant that reads and edits project files, runs shell commands, manages… claude-code-best/claude-code — Claude Code is a command-line interface and multi-agent orchestration framework designed for autonomous software… kunchenguid/lavish-axi — Lavish-axi is a review toolkit for HTML pages that AI coding agents generate. It opens an agent's HTML artifact in a… enzed/vibe-coding — Vibe-coding is an agentic workflow manager and AI coding orchestrator designed to guide autonomous agents through…