{"_id":"@botdigit/agent-blueprint","name":"@botdigit/agent-blueprint","dist-tags":{"latest":"1.0.0"},"versions":{"1.0.0":{"name":"@botdigit/agent-blueprint","version":"1.0.0","description":"The universal blueprint and skill standard for AI coding agents — teaching agents how to think, audit, architect, and properly document any software project.","main":"bin/cli.js","bin":{"agent-blueprint":"bin/cli.js"},"scripts":{"test":"node bin/cli.js status"},"repository":{"type":"git","url":"git+https://github.com/botdigit-official/agent-blueprint.git"},"keywords":["ai-agents","coding-agents","agentic-workflows","cursor","claude-code","antigravity","windsurf","developer-tools","prompt-engineering","software-architecture","code-audit","documentation"],"author":{"name":"BotDigit","url":"https://botdigit.com"},"license":"MIT","bugs":{"url":"https://github.com/botdigit-official/agent-blueprint/issues"},"homepage":"https://github.com/botdigit-official/agent-blueprint#readme","gitHead":"dee11e570a3e974a6b294c72eb30fe8f7b2e20dd","_id":"@botdigit/agent-blueprint@1.0.0","_nodeVersion":"25.8.1","_npmVersion":"11.11.0","dist":{"integrity":"sha512-Z0mbJhRU6EV7CKlSoUk4oZobjZrAc07u5Z6yAXwhzrEkIbjYBcpCxdTzenvcpSx5M+kNECcSqgaVttLVwOGtOA==","shasum":"5bc5a9448f08e3768807214a64dda07dce8087e2","tarball":"https://registry.npmjs.org/@botdigit/agent-blueprint/-/agent-blueprint-1.0.0.tgz","fileCount":52,"unpackedSize":141899,"signatures":[{"keyid":"SHA256:DhQ8wR5APBvFHLF/+Tc+AYvPOdTpcIDqOhxsBHRwC7U","sig":"MEUCIQDiJZvTABEKjcSfcLO6Rw77U3/pNa7GLIV5pe5XgqrsRAIgYWd1gl/6qGxYKo3gAC3MOuJXsv7TmDHVJytfYftpVLE="}]},"_npmUser":{"name":"botdigit","email":"botdigit@gmail.com"},"directories":{},"maintainers":[{"name":"botdigit","email":"botdigit@gmail.com"}],"_npmOperationalInternal":{"host":"s3://npm-registry-packages-npm-production","tmp":"tmp/agent-blueprint_1.0.0_1788944122274_0.6298628157163189"},"_hasShrinkwrap":false}},"time":{"created":"2026-09-09T08:55:22.059Z","1.0.0":"2026-09-09T08:55:22.427Z","modified":"2026-09-09T08:55:22.655Z"},"maintainers":[{"name":"botdigit","email":"botdigit@gmail.com"}],"description":"The universal blueprint and skill standard for AI coding agents — teaching agents how to think, audit, architect, and properly document any software project.","homepage":"https://github.com/botdigit-official/agent-blueprint#readme","keywords":["ai-agents","coding-agents","agentic-workflows","cursor","claude-code","antigravity","windsurf","developer-tools","prompt-engineering","software-architecture","code-audit","documentation"],"repository":{"type":"git","url":"git+https://github.com/botdigit-official/agent-blueprint.git"},"author":{"name":"BotDigit","url":"https://botdigit.com"},"bugs":{"url":"https://github.com/botdigit-official/agent-blueprint/issues"},"license":"MIT","readme":"<div align=\"center\">\n\n# 📐 Agent Blueprint\n\n**The universal blueprint and skill standard for AI coding agents — teaching agents how to think, audit, architect, and properly document any software project.**\n\n[![CI](https://github.com/botdigit-official/agent-blueprint/actions/workflows/ci.yml/badge.svg)](https://github.com/botdigit-official/agent-blueprint/actions)\n[![License: MIT](https://img.shields.io/badge/License-MIT-blue.svg)](LICENSE)\n[![PRs Welcome](https://img.shields.io/badge/PRs-welcome-brightgreen.svg)](CONTRIBUTING.md)\n[![Agents](https://img.shields.io/badge/Agents-Antigravity%20%7C%20Claude%20Code%20%7C%20Cursor%20%7C%20Windsurf%20%7C%20Cline-orange)](#)\n[![GitHub Stars](https://img.shields.io/github/stars/botdigit-official/agent-blueprint?style=social)](https://github.com/botdigit-official/agent-blueprint)\n\n<br/>\n\n> *\"The agent must adapt to the project. The project must not be forced to adapt to the skill.\"*\n\n</div>\n\n---\n\n## ⚡ 1-Minute Quickstart\n\nInstall **Agent Blueprint** into any project with your preferred method:\n\n### Option A: One-Liner (Recommended)\n```bash\ncurl -fsSL https://raw.githubusercontent.com/botdigit-official/agent-blueprint/main/install.sh | bash\n```\n\n### Option B: Via npx / Node\n```bash\nnpx agent-blueprint init\n```\n\n### Option C: Git Clone\n```bash\ngit clone --depth 1 https://github.com/botdigit-official/agent-blueprint.git ~/.agent-blueprint\n~/.agent-blueprint/install.sh .\n```\n\n### 🤖 Multi-Agent Compatibility Out of the Box\n| AI Tool | Configuration File Generated | Workflow Activated |\n|---|---|---|\n| **Google Antigravity** | `AGENTS.md` + `.agents/skills/` | Full modular skill hierarchy |\n| **Claude Code** | `CLAUDE.md` | Context & architecture guardrails |\n| **Cursor AI** | `.cursorrules` | Architectural & testing rules |\n| **Windsurf / Cline / Aider** | `AGENTS.md` | Autonomous discovery & audit |\n\n---\n\n## What is Agent Blueprint?\n\nMost AI coding agents jump straight to generating code or rewriting working systems without understanding the domain, architecture, dependencies, or conventions.\n\n**Agent Blueprint** is a composable, technology-agnostic skill and documentation standard. It provides structured guidance so any AI agent can:\n\n1. **Discover** the project layout and entry points across the repo\n2. **Understand** its stack, framework, database, and maturity\n3. **Audit** what exists before modifying or breaking anything\n4. **Document** findings into living, structured architectural documentation\n5. **Architect** changes grounded in real business needs, not guesswork\n6. **Plan** features with Architectural Decision Records (ADRs)\n7. **Implement** following existing patterns and conventions\n8. **Test & Verify** to preserve regression safety and data integrity\n9. **Review & Audit** against security and performance baselines\n\n> 🤖 **Direct LLM Ingestion**: Feed [**`llms.txt`**](llms.txt) to web crawlers/agents or copy the 1-click system prompt in [**`PROMPT.md`**](PROMPT.md).\n\n---\n\n## ⚖️ Why You Need This: Raw AI vs. Agent Blueprint\n\n| Scenario | Raw AI / Default Copilot | With Agent Blueprint |\n|---|---|---|\n| **Entering a project** | Guesses architecture, invents new dependencies | Runs discovery, detects existing stack, checks `docs/` |\n| **Refactoring code** | Often rewrites working systems and breaks logic | Obeys established ADRs, respects existing state machines |\n| **Documentation** | Leaves zero comments and zero docs | Automatically updates `docs/` living documentation |\n| **New Features** | Injects unverified patterns | Writes automated tests, updates API specs |\n| **Multi-Agent Teams** | Claude Code, Cursor, and Antigravity fight | All agents share a single source of truth |\n\n---\n\n## Core Philosophy\n\nThese rules live at the top of every skill and guide every agent decision:\n\n1. **Understand before changing.**\n2. **Inspect before assuming.**\n3. **Document before redesigning.**\n4. **Prefer simple architecture.**\n5. **Use the project's existing technology when practical.**\n6. **Do not introduce technology without a measurable reason.**\n7. **Business logic comes before code structure.**\n8. **Security and data integrity come before features.**\n9. **Deterministic tools before AI.**\n10. **AI only where reasoning adds value.**\n11. **APIs only where they provide unique value.**\n12. **Never rewrite working systems unnecessarily.**\n13. **Preserve existing functionality unless explicitly deprecated.**\n14. **Every architectural decision needs a reason (ADR).**\n15. **Every major change needs automated tests.**\n16. **Every project needs a source of truth.**\n17. **Every agent must leave the project better documented than it found it.**\n\n---\n\n## How It Works\n\n### Composable Skills, Not a Giant System Prompt\n\nAgent Blueprint is **not** a single unwieldy prompt. It is a directory of modular, composable skills. The autonomous orchestrator selects only what applies to the current project:\n\n```\nskills/\n├── 00-orchestrator/           # Triage, detect, activate\n├── 01-discovery/              # Find project roots in monorepos or nested dirs\n├── 02-project-context/        # Stack, framework, database, infrastructure\n├── 03-business-architecture/  # Domain, actors, business rules, workflows\n├── 04-architecture/           # Architectural review, ADRs, gaps\n├── 05-documentation/          # Living documentation standard\n├── 06-codebase-audit/         # Code-level inspection and hygiene\n├── 07-security/               # Security review, secrets, CSRF, auth audit\n├── 08-testing/                # Test strategy, coverage, regression suite\n├── 09-performance/            # Latency, queries, throughput, bottlenecks\n└── 10-audit/                  # Combined forensic audit workflow\n```\n\nStack & framework adapters activate automatically based on detected project manifests (`Cargo.toml`, `package.json`, `pyproject.toml`, `go.mod`, `pom.xml`, etc.):\n\n```\nstacks/                frameworks/\n├── rust/              ├── axum/\n├── typescript/        ├── nextjs/\n├── python/            ├── react/\n├── go/                ├── fastapi/\n├── php/               ├── django/\n├── java/              ├── laravel/\n├── ruby/              ├── spring/\n└── dotnet/            └── rails/\n```\n\n### Minimum Sufficient Documentation\n\nNot every project needs 50 documents. Agent Blueprint selects the minimum documentation set appropriate to the project's size and risk:\n\n| Project Scope | Required Documentation Baseline |\n|---|---|\n| **Script / CLI Tool** | `README.md` only |\n| **Web App / API** | Project Brief, Architecture, Database, API Spec, Security |\n| **SaaS / Marketplace** | + Business Architecture, Testing, Deployment, Runbook |\n| **Regulated / Enterprise** | + Compliance, Audit Trail, Data Governance, Disaster Recovery |\n\n---\n\n## Repository Structure\n\n```\nagent-blueprint/\n├── README.md                  # Project overview & quickstart\n├── AGENTS.md                  # Primary instruction manual for coding agents\n├── CONTRIBUTING.md            # Guidelines for community skill additions\n├── LICENSE                    # MIT License\n├── install.sh                 # 1-click project linker & installer\n│\n├── skills/                    # Core, technology-independent skills\n│   ├── 00-orchestrator/\n│   ├── 01-discovery/\n│   ├── 02-project-context/\n│   ├── 03-business-architecture/\n│   ├── 04-architecture/\n│   ├── 05-documentation/\n│   ├── 06-codebase-audit/\n│   ├── 07-security/\n│   ├── 08-testing/\n│   ├── 09-performance/\n│   └── 10-audit/\n│\n├── stacks/                    # Technology adapters (Rust, TS, Python, Go, etc.)\n├── frameworks/                # Framework adapters (Next.js, Axum, FastAPI, etc.)\n│\n├── templates/                 # Production-grade documentation templates\n│   ├── project-brief/         # Initial scoping & objectives\n│   ├── business-requirements/ # Domain entities & actor rules\n│   ├── architecture/          # System design & component diagrams\n│   ├── adr/                   # Architecture Decision Records\n│   ├── api-spec/              # REST / GraphQL API contracts\n│   ├── database/              # Schema, migrations & relationships\n│   ├── security/              # Threat models & access controls\n│   ├── testing/               # Quality gates & verification plans\n│   ├── deployment/            # CI/CD pipelines & hosting\n│   ├── runbook/               # Incident response & operational SOPs\n│   └── changelog/             # Semantic version releases\n│\n└── examples/                  # Real-world audits & walkthroughs\n    ├── botdigit-site/         # Multi-tenant directory & site builder\n    ├── saas/                  # Multi-tier subscription platform\n    ├── marketplace/           # Two-sided buyer/seller marketplace\n    ├── fintech/               # High-security payment processing\n    ├── ecommerce/             # Catalog, cart, and order fulfillment\n    ├── directory/             # Geo-spatial search & discovery\n    ├── mobile-app/            # Cross-platform iOS/Android app\n    ├── ai-product/            # LLM agents & deterministic pipelines\n    └── internal-tool/         # Backoffice admin control panels\n```\n\n---\n\n## Quick Start\n\n### For AI Coding Agents\nRead [`AGENTS.md`](AGENTS.md) first. It specifies the step-by-step discovery, inspection, and execution lifecycle.\n\n### For Developers & Tech Leads\n1. Browse [`skills/`](skills/) to understand the reasoning frameworks.\n2. Check [`templates/`](templates/) for ready-to-use architecture and documentation templates.\n3. Review [`examples/botdigit-site/`](examples/botdigit-site/) to see a real forensic audit and remediation.\n\n---\n\n## Contributing\n\nWe welcome contributions! You can add new technology stacks, framework adapters, or domain patterns:\n1. Review [`CONTRIBUTING.md`](CONTRIBUTING.md).\n2. Follow the metadata schema in `skills/00-orchestrator/SKILL.md`.\n3. Submit a Pull Request.\n\n---\n\n## Frequently Asked Questions (FAQ)\n\n### How does Agent Blueprint prevent AI from breaking existing code?\nAgent Blueprint enforces a strict rule: **\"The agent must adapt to the project; the project must not adapt to the skill.\"** Before an agent is allowed to write or edit code, it must execute Discovery, inspect existing frameworks, review active state machines, and check existing unit tests.\n\n### Does this work with Cursor, Claude Code, and Antigravity?\n**Yes.** When you run `./install.sh` or `npx agent-blueprint init`, it automatically creates `.cursorrules` (for Cursor), `CLAUDE.md` (for Claude Code), and `AGENTS.md` + `.agents/skills/` (for Antigravity, Cline, Windsurf). All agents follow the exact same architectural guidelines.\n\n### What is \"Living Documentation\"?\nInstead of outdated wikis or empty READMEs, Agent Blueprint establishes a structured `docs/` hierarchy (Business Model, Architecture Decision Records, Database Schemas, API Specs, Security). Every time an agent modifies system behavior, it is required to update the corresponding document.\n\n### How is this different from a system prompt?\nA single giant prompt gets truncated and forgotten in long context windows. Agent Blueprint is a **modular, composable skill hierarchy**. The autonomous orchestrator activates only the skills required for the specific task at hand.\n\n---\n\n## License\n\nMIT © [BotDigit](https://botdigit.com) — see [LICENSE](LICENSE).\n\n","readmeFilename":"README.md","_rev":"1-1ffaf438d44e67e7f44ac112c47d9752"}