{"name":"AIWG: Reusable Context & Workflows for AI-Augmented Development","description":"AIWG is a cognitive architecture designed to enhance AI-augmented software development. It provides reusable project context and specialist workflows, enabling structured development, review, and operational tasks across various AI tools and platforms.","github":"https://github.com/jmagly/aiwg","url":"https://osrepos.com/repo/jmagly-aiwg","source":"osrepos.com","sourceDescription":"This repository profile is provided by osrepos.com, an open source repository discovery platform.","repositoryProfile":"https://osrepos.com/repo/jmagly-aiwg","generatedFor":"open source discovery and AI-assisted research","markdown":"https://osrepos.com/repo/jmagly-aiwg.md","json":"https://osrepos.com/repo/jmagly-aiwg.json","topics":["ai-framework","multi-agent","agentic-coding","sdlc","workflow-automation","developer-tools","typescript","orchestration"],"keywords":["ai-framework","multi-agent","agentic-coding","sdlc","workflow-automation","developer-tools","typescript","orchestration"],"stars":null,"summary":"AIWG is a cognitive architecture designed to enhance AI-augmented software development. It provides reusable project context and specialist workflows, enabling structured development, review, and operational tasks across various AI tools and platforms.","content":"## Introduction\n\nAIWG is an innovative multi-agent AI framework designed to augment software development and various other complex tasks. It provides reusable project context and specialist workflows for the AI tools you already use, ensuring consistency and efficiency across sessions. With specialized agents, structured workflows, and multi-platform deployment, AIWG helps teams plan software, coordinate reviews, manage incidents, and maintain operational knowledge. It integrates with popular AI providers like Claude Code, Codex, Copilot, and Cursor, offering a unified approach to AI-augmented work.\n\n## Why Use AIWG and Key Benefits\n\nAIWG addresses common challenges in AI-assisted projects, such as maintaining context, recovering from failures, and making quality criteria explicit.\n\n### Problems AIWG Solves:\n\n*   **Maintaining Context Across Sessions:** Useful decisions often get scattered. AIWG workflows save project outputs in the `.aiwg/` directory, including requirements, architecture decisions, risk notes, and test strategies, making them retrievable for later tasks.\n*   **Recovering from Failed Attempts:** Instead of just retrying, AIWG's agent loops support an execute-and-verify cycle. This records failure information, adapts the next attempt, and stops at configured limits, making the recovery process reviewable.\n*   **Making Quality Criteria Explicit:** AIWG provides specialist roles and workflows that separate concerns like security, performance, and testing. These roles combine findings and record open decisions, ensuring required artifacts and reviews are ready before work advances.\n\n### Six Core Components:\n\nAIWG is built on simple, yet powerful, building blocks that compound in usefulness:\n\n1.  **Memory, Structured Semantic Memory:** The `.aiwg/` directory acts as a persistent artifact repository for project knowledge, grounding later work in recorded decisions.\n2.  **Reasoning, Multi-Agent Deliberation with Synthesis:** Specialist role definitions, like Security Auditor or Test Architect, allow complex artifacts to be routed through required reviewers, with synthesis to merge feedback.\n3.  **Learning, Closed-Loop Self-Correction (Agent Loop):** Executes tasks iteratively, using verification results to guide subsequent attempts and preserving failure analysis.\n4.  **Verification, Bidirectional Traceability:** Supports links between documentation and code, allowing reviewers to inspect relationships and find drift, for example, `@implements @.aiwg/requirements/UC-001-login.md`.\n5.  **Planning, Phase Gates with Cognitive Load Management:** Structures work using a Stage-Gate methodology, breaking projects into bounded phases with explicit quality criteria and human approval.\n6.  **Style, Controllable Voice Generation:** Voice profiles describe writing preferences, such as formality or technical depth, providing reusable style specifications for content generation.\n\nAIWG also offers a rich catalog of frameworks, including SDLC Complete, Forensics Complete, Media/Marketing Kit, and Research Complete, along with numerous addons like Agent Loop, RLM (Recursive Context Decomposition), and Compound Memory, extending its capabilities across various domains.\n\n## Installation\n\nAIWG offers both a prompt-led installer and manual installation via npm.\n\n### Prerequisites:\n\nEnsure you have Node.js >=20.0.0 and an AI platform (e.g., Claude Code, GitHub Copilot, Cursor, Warp Terminal).\n\n### Prompt-Led Installation (Recommended):\n\nPaste the following into a supported AI provider:\n\ntext\nInstall or repair AIWG for this project by following\nhttps://aiwg.io/setup.aiwg.yaml\nExplain the plan before changing anything, preserve my existing work, and ask\nme only for choices you cannot safely determine.\n\n\n### Manual Installation:\n\nFor a manual setup, use npm:\n\nbash\nnpm i -g aiwg\ncd /path/to/your/project\naiwg use all --provider claude   # Replace 'claude' with your AI provider selector\n\n\nAfter installation, you can check the deployment health and readiness with:\n\nbash\naiwg doctor\n\n\nYou can also deploy specific frameworks instead of `all`, for example:\n\nbash\naiwg use sdlc --provider claude          # Software development lifecycle\naiwg use forensics --provider claude     # Investigation workflows\n\n\n## Examples\n\nHere are a few ways to leverage AIWG in your projects:\n\n### Get a First Useful Result:\n\nAfter setup, ask your agent to review documentation:\n\ntext\nUse AIWG to review this project's README for unclear positioning and missing\nonboarding steps. Save a report at\n.aiwg/marketing/brand/audit/readme-review.md with file references and the\nthree highest-priority fixes. Leave the README unchanged.\n\n\n### SDLC Workflow from Idea to Implementation:\n\nInitiate a comprehensive development lifecycle:\n\ntext\nTurn an AI-powered code review tool into requirements, architecture decisions, a\nfirst implementation task, and acceptance checks. Return links to the generated artifacts.\n\n\n### Long-Running Implementation Loop:\n\nFix issues iteratively with defined boundaries:\n\ntext\nFix the failing authentication tests in a bounded loop. Stop after five iterations or forty-five minutes, and\nreport whether the relevant tests pass and what remains unresolved.\n\n\n## Links\n\n*   **GitHub Repository:** [jmagly/aiwg](https://github.com/jmagly/aiwg){target=\"_blank\"}\n*   **Official Website:** [aiwg.io](https://aiwg.io){target=\"_blank\"}\n*   **Discord Community:** [Join Server](https://discord.gg/BuAusFMxdA){target=\"_blank\"}\n*   **Telegram Group:** [Join Group](https://t.me/+oJg9w2lE6A5lOGFh){target=\"_blank\"}\n*   **Documentation:** [AIWG Docs](https://aiwg.io/docs){target=\"_blank\"}","metrics":{"detailViews":0,"githubClicks":0},"dates":{"published":null,"modified":"2026-09-19T15:47:15.000Z"}}