Ai Scrum Helper

Scrum Master AI Helper

Experimental. A markdown-based, assistant-agnostic AI agent that uses MCP (Jira and Confluence) to automate sprint health checks, onboarding, and stakeholder summaries. No code—only natural language instructions and playbooks.

What it does

• Sprint health (/health): Stale tickets (no movement in 3+ days), scope creep, unpointed work, and recommendations.
• Sprint summary (/summary): Stakeholder-ready review of completed work, grouped by theme, with business-value language.
• Onboarding (/onboard): Discovers Jira/Confluence config (project, sprint, custom fields, team norms) and writes scrum-master/PROJECT_CONTEXT.md.

Quick start

1. MCP: Use an MCP server that can talk to Jira (and optionally Confluence). For example, the Atlassian HTTP MCP at https://mcp.atlassian.com/v1/mcp.
2. Assistant: Use Cursor or another AI assistant that can run MCP tools and follow the playbooks.
3. Run /onboard once to create PROJECT_CONTEXT.md, then use /health and /summary as needed.

Commands

| Command | Purpose | |------------|---------| | /onboard | Discover and save Jira/Confluence config to scrum-master/PROJECT_CONTEXT.md. Run first or when project/sprint changes. | | /health | Analyze active sprint: stale tickets, scope creep, unpointed work; output markdown report and recommendations. | | /summary | Generate sprint review summary (Done work, themes, metrics) for stakeholders. |

Layout

• scrum-master/playbooks/ – Step-by-step instructions (onboard, sprint-health, review-summary) that the AI follows.
• scrum-master/AGENTS.md – Persona and behavior (proactive, business language, actionable).
• scrum-master/PROJECT_CONTEXT.md – Created by /onboard; gitignored. Source of truth for project key, sprint, custom fields, team norms.
• .claude/commands/ and .cursor/rules/ – Wire slash commands and rules to the playbooks.

Configuration

• Cursor: Rules in .cursor/rules/index.mdc point to PROJECT_CONTEXT.md and the playbooks.
• Claude Code: Slash commands in .claude/commands/ invoke the same playbooks.
• Other assistants: Use CLAUDE.md, scrum-master/playbooks/, and scrum-master/AGENTS.md; config lives in PROJECT_CONTEXT.md.

Extending

• New command: Add a playbook under scrum-master/playbooks/, then add a command/rule that references it.
• New tools: Playbooks are written in semantic terms (“fetch sprint issues”, “search Confluence”); adapt them to whatever MCP or API your Jira/Confluence integration exposes.

Troubleshooting

• No active sprint / PROJECT_CONTEXT missing: Run /onboard. Ensure an active sprint and board exist in Jira.
• Confluence empty: Optional. You can fill team norms manually in PROJECT_CONTEXT.md.
• Custom fields: Playbooks assume common fields (e.g. story points, sprint). Map your field IDs in PROJECT_CONTEXT.md; /onboard tries to discover them.

Disclaimer

This is a personal, open-source project built independently. It does not include or rely on any confidential or proprietary information.

License

MIT. See [LICENSE](LICENSE.md) or repository root for full text.

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

• Author: apetcu
• Source: apetcu/ai-scrum-helper
• License: Apache-2.0

Install and usage instructions live in the source repository linked above.