This project is a workshop proof-of-concept showing how AI can automate and improve the release note workflow.
The project uses release notes as the business use case to demonstrate Claude capabilities such as commands, skills, agents, marketplace concepts, plugins, and MCP.
Demonstrate an end-to-end AI-assisted workflow:
Change information → AI extraction → Release note draft → AI review → Improved draft → Human approval
- charter: Project scope, goals, success criteria
- roster: Team members and roles
- samples: Sample input and output files
- prompts-and-commands: Reusable prompts and commands
- skills: Reusable skill definitions or instructions
- agent-workflow: Multi-step AI workflow design
- mcp-plugin-concept: Integration ideas using MCP, plugins, or marketplace tools
- validation: Test results, risks, limitations
- demo: Final demo material
- meetings: Meeting notes and action items
- Clone repository
- Read charter
- Review roster
- Select assigned tasks (in team-roster.md)
- Update work in only this repository
- Raise issues when blocked
The guideline explicitly says:
Note No critical work should live only on a personal machine, private drive, or chat thread.
Therefore: ❌ Don't allow: "I have the prompt on my laptop." "I shared it in WhatsApp." "I emailed it to someone."
✅ Instead use:
- prompts-and-commands/
- skills/
- validation/
- demo/
Everything gets committed to the repo.
Success means: Demonstrated a realistic documentation workflow
- Showed Commands, Skills, and Agents
- Explained MCP and integrations
- Used realistic business scenarios
- Demonstrated measurable workflow improvement
- Identified human review checkpoints
We are not building a release note tool. We are demonstrating how AI capabilities can automate a documentation workflow using release notes as the business use case.