Pudo

Enhance npm package publishing and documentation for MCP server

PUDO just got a major upgrade: it’s now more than a prompt library. The repo includes an installable AI agent operating layer with CLI commands, quality gates, repo-level rules for Cursor, Claude, Codex, Copilot, Gemini, OpenCode, and Kiro, plus context engineering docs, benchmark kit, and CI checks. Built to help teams move from vibe coding to structured, measurable AI-assisted development.

Added support for Gemini, OpenCode, and Kiro AI coding agents in PUDO Code. The CLI now recognizes these agents during init, config, scoring, and evaluation, and generates their instruction files automatically. Docs, package files, and tests were updated to cover the new agent workflows.

PUDO now has a real CLI test suite powered by Node’s built-in test runner, covering argument parsing, template generation, dry-run, force overwrite, and check/score/doctor flows. I also added GitHub Actions CI for push and PR validation. This makes PUDO more reliable for AI-assisted coding workflows with Codex, Claude, Cursor, and Copilot.

- pudo analyze, schema validation, report JSON/Markdown. - AI stack presets + real examples cho Next.js, FastAPI, RAG app.

- Repositioned PUDO as an AI Agent Operating Layer across Cursor, Claude, Codex, GitHub Copilot, and Gemini/Antigravity - Synchronized package and README badge version to 1.1.0

- Added Expected Impact: ~34% less wasted AI tokens, ~18% faster dev time. - Improved PUDO workflow for real AI-assisted coding: Plan → Understand → Develop → Optimize. - Added stronger phase rules, quality gates, AI output review, token-aware guidance, and multi-agent integrations.

- Added modern AI agent config support for Claude Code, Cursor, Copilot, Gemini, and Codex. - Migrated Cursor guidance from `.cursorrules` to `.cursor/rules/*.mdc`. - Added Copilot, Claude, and Gemini instruction files. - Improved Codex `AGENTS.md` with task modes, context budgets, patch discipline, and verification tiers. - Fixed outdated Claude README references. - Clarified PUDO depth for micro, standard, and critical tasks. - Added recommended Markdown/link/instruction checks.

- Added Micro, Standard, and Critical task modes. - Improved context usage with targeted search and focused file inspection. - Limited unnecessary file reading and large output dumps. - Added patch discipline to prevent unrelated rewrites. - Added verification tiers for faster, task-appropriate checks. - Standardized concise final responses: Changed, Verified, Notes. - Preserved PUDO while making Codex faster and leaner.

Big milestone: PUDO Code System just became the #1 Launch Batch Winner on Forg. I’ve also recently updated the workflow to work better across all AI agents, making it easier for builders to structure, rewrite, and improve their projects with more consistency. If you’re currently building something, feel free to try it here: https://github.com/DongDuong2001/pudo-code-system Would love to see what people build with it.

Quality Gates, QC checklists, edge-case libraries, AI output review, failure-mode tracking, Definition of Done templates, PR checklist, and release discipline. This update turns PUDO from a prompt methodology into a structured AI engineering workflow for safer, more reliable coding with AI.

- Update Rule for Codex. - Optimize Workflow avoid unecessary token burning.

I'm excited to shared that my PUDO Github Repo just hit #2 product of the Week on Unikorn Upvote on Unikorn: https://unikorn.vn/p/pudo Thank you everyone!!!

• Next.js 15 — RSC, Server Actions, caching strategies • Python FastAPI — Pydantic v2, async/await rules • Rust Backend — lifetimes, Axum/Actix, cargo optimization • Web3 Smart Contracts — CEI patterns, gas optimization, security • React Native Expo — Reanimated, FlatList, UI threading

I’m currently adding support for the PUDO code system project. PUDO turns prompt-based development into a structured and repeatable system, helping developers organize workflows, standardize outputs, and improve maintainability for AI-assisted projects. Current work includes integration support, workflow handling, and compatibility testing to make PUDO reliable and scalable. Support the project: https://github.com/sponsors/DongDuong2001

Happy to share that this workflow repo just hit #2 Product of the Day on Unikorn Come check it out: https://unikorn.vn/p/pudo

PUDO is a system of agent skills and structured prompts that helps developers build software with more consistency and control when working with AI. Built from real project experience, it organizes AI usage into reusable workflows for planning, understanding, developing, and optimizing systems which reducing random outputs and making development more scalable and maintainable.