diff --git a/README.md b/README.md index 54c979644..346495682 100644 --- a/README.md +++ b/README.md @@ -16,8 +16,11 @@ ## Table of Contents - [Overview](#overview) -- [How JUNO Was Born](./docs/history.md) -- [Why JUNO is Agentic AI, Not Just AI Agents](#why-juno-is-agentic-ai-not-just-ai-agents) +- [Problem Statement & Solution Architecture](#problem-statement--solution-architecture) +- [Phase-Based Agentic AI Maturity Model](#phase-based-agentic-ai-maturity-model) +- [Why This Architecture Matters](#why-this-architecture-matters) +- [Strategic Outcome](#strategic-outcome) +- [How to Add These Sections in Your README.md](#how-to-add-these-sections-in-your-readmemd) - [Architecture](#architecture) - [Microservices Architecture](#microservices-architecture) - [Technology Stack](#technology-stack) @@ -81,23 +84,105 @@ --- -## How JUNO Was Born +## 🧠 Problem Statement & Solution Architecture -For a detailed history of the project, consult [docs/history.md](./docs/history.md). +**The Problem: Jira Tracks—But It Doesn’t Think** -## Why JUNO is Agentic AI, Not Just AI Agents +Engineering teams rely on Jira to track sprints, defects, and delivery metrics. But as systems scale, Jira becomes a passive ledger, not a reasoning partner. Teams are burdened with chasing down failure patterns across environments, dashboards, and tools. -**[📖 Essential Reading: AI Agents vs Agentic AI Educational Guide](./docs/guides/ai-agents-vs-agentic-ai.md)** +Common breakdowns: +- Sprint retros take hours to synthesize from Jira exports +- Velocity stalls traced to defects—but root causes remain unclear +- Test failures are logged but not categorized across test data, environment (NPE), script quality, or tech debt +- Engineering leaders drown in dashboards but lack decision-ready insight -JUNO represents true **agentic AI** - autonomous systems that reason through multiple valid paths to achieve optimal outcomes. Unlike traditional AI agents that follow fixed workflows, JUNO's agentic approach enables: +Despite Jira’s extensibility, it delivers information—not understanding. -- **Multiple Valid Solution Paths**: JUNO can resolve sprint risks through reassignment, scope adjustment, or stakeholder escalation - choosing the optimal approach based on context -- **Proactive Intelligence**: Predicts and prevents issues rather than just responding to them -- **Autonomous Reasoning**: Makes informed decisions within defined boundaries without constant human oversight -- **Self-Optimizing**: Learns from outcomes to improve future decision-making -- **Multi-Agent Orchestration**: Coordinates specialized agents for complex workflow management +**The Solution: JUNO as an Agentic AI Analyst** -**For Engineers**: Understanding this distinction is critical for implementing, evaluating, and scaling JUNO effectively. [Read the full guide →](./docs/guides/ai-agents-vs-agentic-ai.md) +JUNO transforms Jira into a vertical AI system that doesn’t just summarize data—it reasons through it. Powered by Enterprise GPT, JUNO performs multi-dimensional defect analysis across: +- Test Script Failures (broken automation logic, brittle assertions) +- Test Data Gaps (expired or missing synthetic records) +- Non-Prod Environment (NPE) Instability (lab-specific defects) +- Structural Tech Debt (recurring code smells or legacy gaps) + +Instead of manual categorization and root-cause hunting, teams ask: + +“Why did regression failures spike last sprint?” +“Which NPE is introducing the most flakiness?” +“Are stale test scripts slowing velocity?” + +JUNO parses Jira exports, applies reasoning, and responds with correlated insights, visual trends, and defensible recommendations. + +--- + +## 🧭 Phase-Based Agentic AI Maturity Model + +JUNO’s development follows a modular framework rooted in agentic AI design: memory, autonomy, reasoning, and observability. + +| Phase | Objective | Key Capabilities | Agentic Alignment | +| ----- | --------- | ---------------- | ---------------- | +| **Phase 1: Analytics Foundation** | Summarize and structure Jira data | Natural language queries, sprint metrics, defect heatmaps | 📊 Insight Delivery | +| **Phase 2: Agentic Workflow Management** | Reason about blockers and delivery risk | Risk forecasts, memory layers, test defect diagnostics | 🧠 Autonomous Reasoning + Episodic Memory | +| **Phase 3: Multi-Agent Orchestration** | Align insights across squads and platforms | Coordination agents, consensus, fault recovery | 🔁 Distributed Cognition | +| **Phase 4: AI-Native Operations** | Predict and prevent delivery failure | RL optimization, anomaly detection, self-healing logic | ⚙ Autonomy at Scale | + +--- + +## đŸ§© Why This Architecture Matters + +JUNO adheres to enterprise-grade AI standards: +- Memory Hierarchies: episodic (per sprint), semantic (per workflow), procedural (per test) +- Transparent Reasoning: confidence scores, traceable audits +- Governance: role-based approval, secure data flow +- Observability: latency metrics, defect category accuracy, risk deltas + +It doesn’t just categorize failure—it understands it. + +--- + +## 📐 Strategic Outcome + +JUNO closes the gap between defect logging and engineering intelligence. It transforms Jira into a decision engine that reduces risk, accelerates retros, and clarifies velocity blockers at scale. + +It’s not another Jira app. It’s the analyst we needed—but could never hire. + +--- + +## ✅ How to Add These Sections in Your README.md + +GitHub doesn’t support these natively in raw Markdown, but you can simulate them using blockquotes: + +> **Note** +> OAuth 2.0 is more complex to set up but provides enhanced security features. For most users, API Token authentication (Method A) is simpler and sufficient. + +> **Important** +> Include `offline_access` in scope for persistent auth (e.g., `read:jira-work write:jira-work offline_access`) + +### 📩 Installation + +MCP Atlassian is distributed as a Docker image. This is the recommended way... + +> 💡 **Tip** +> For Claude Desktop: Locate and edit the configuration file directly: + +--- + +### đŸ§Ș Bonus: Custom Styling (if hosted elsewhere) + +If you’re using a framework like Docusaurus, MkDocs, or Docsify, they often support: + +:::note +This is a note box +::: + +:::tip +Helpful tip goes here. +::: + +:::warning +Something important! +::: ---