npc-guide
v0.2.1
Published
Your repo remembers everything. Silent mission system for AI coding agents.
Maintainers
Readme
npc-guide
Your repo remembers everything. Your AI never starts from zero.
Silent mission system that turns AI coding agents (Claude Code, Cursor, Copilot) into autonomous executors. Install it, give it a brief, and your agent stops asking questions and starts building.
Quick Start
# Install (silent — creates .ai-guide/, CLAUDE.md, hooks automatically)
npm install https://github.com/Abhipaddy8/npc-guide-ai/releases/download/v0.1.0/npc-guide-0.1.0.tgz
# Initialize with your brief
npx npc-guide init "Build a SaaS dashboard with Next.js and Supabase"
# Open your coding agent and type "go"That's it. The agent reads the mission map and executes autonomously.
What It Does
- Silent install — postinstall creates
.ai-guide/, wires Claude Code hooks, writesCLAUDE.mdand.cursorrules - Brief parsing — detects intent (build, strategy, research, debug, etc.) and generates a mission map
- Agent direction — your agent reads the missions and executes without asking questions
- Session memory — every session is archived, every decision logged, agent picks up where it left off
7 Intent Types
| Intent | Mission Sequence | |---|---| | build | Foundation → Core Loop → Data → Auth → UI → Integration → Ship | | strategy | Landscape → Positioning → Growth → Retention → Monetization → Launch | | research | Scope → Gather → Analyze → Synthesize → Recommend | | design | Requirements → System Map → Interface → Validation | | content | Outline → Draft → Polish → Distribute | | ops | Audit → Implement → Test → Rollout | | debug | Reproduce → Isolate → Fix → Verify |
Agent Support
| Agent | Integration | |---|---| | Claude Code | SessionStart/End hooks + CLAUDE.md (full lifecycle) | | Cursor | .cursorrules | | Copilot / Others | CLAUDE.md (any agent that reads project markdown) |
No API Required
Default parser is pure regex — works offline, instantly. LLM parsing (OpenAI/Anthropic/OpenRouter) is optional for richer results.
Docs
Full documentation, architecture, and live test results: github.com/Abhipaddy8/npc-guide-ai
License
MIT
