Agentic SDLC Framework

This repository is an agent-first software development lifecycle framework designed to build self-enforcing, production-ready multi-agent systems. Every agent operating within this repository MUST load and internalize principles/agentic-playbook/index.md before initiating any reasoning or tool use.

Features & Tooling

• LLM-as-Judge CI/CD: PRs and commits are automatically evaluated against the 15 Agentic Principles by a GitHub Actions workflow (.github/workflows/agent-principle-check.yml).
• Reusable Harness: The Bootstrap Kit allows you to apply this agentic SDLC framework and git hooks to any existing repository.
• Agent CLI Tool: Use sdlc-cli.py to quickly verify compliance and load principles.
• IDE & Framework Integrations: Includes automated guidelines for Cursor IDE (.cursor/rules/sdlc.mdc) and an example CrewAI dynamic skills loader (50-examples/integrations/crewai_skills_loader.py).

How Agents Use This Repo

• Mandatory Initialization: Every agent session must begin by loading the Agent Bootstrap Prompt and reading the core principles.
• Workflow Alignment: Agents design and execute tasks in strict accordance with the canonical rules established in the repository blueprint.
• Continuous Self-Checking: Agents must continuously self-evaluate their reasoning traces against the non-negotiable principles before taking destructive actions.
• Template Utilization: Agents rely on the provided templates to ensure uniform output, comprehensive testing, and standardized PRs.
• Context Awareness: Agents consult context files to maintain awareness of repository boundaries, tools, and environmental constraints.

Key Resources

• Agentic Principles Playbook
• Agentic SDLC Blueprint
• Repository Context
• Agent Guide & Contributing Guide

Bootstrap Command

To initialize an agent session, apply the following prompt: cat 40-templates/agent-bootstrap-prompt.md