Github Project Management

GitHub Project Management

Coordinates GitHub issue management, project board automation, and sprint planning using AI swarm agents. Supports issue triage / decomposition, bidirectional board sync, automated progress tracking, and KPI reporting.

Contents

• When to Use
• Prerequisites
• Quick Start
• Workflow
• Verification
• References

When to Use

• "Triage these unlabelled issues"
• "Break down this epic into subtasks"
• "Sync GitHub issues with the project board"
• "Set up a sprint with velocity tracking"
• "Generate a sprint-summary report"
• "Coordinate work across multiple repos / boards"

Prerequisites

See references/troubleshooting.md for full tool versions, required token scopes, and GitHub API limits.

Minimum to run the plain gh examples: gh + jq. Swarm features additionally need ruv-swarm or claude-flow MCP server running — see those projects' docs for install. The skill degrades gracefully: anything prefixed with npx ruv-swarm requires the MCP server; everything else is plain gh + jq.

Quick Start

# Create a coordinated issue
gh issue create \
  --title "Feature: Advanced Authentication" \
  --body "Implement OAuth2 with social login..." \
  --label "enhancement,swarm-ready"

# Initialise swarm tracking
npx claude-flow@alpha hooks pre-task --description "Feature implementation"

# Set up the board
PROJECT_ID=$(gh project list --owner @me --format json | jq -r '.projects[0].id')

npx ruv-swarm github board-init \
  --project-id "$PROJECT_ID" \
  --sync-mode "bidirectional"

Workflow

1. Triage incoming issues — auto-label, link duplicates (issue-management.md)
2. Decompose large issues into subtasks with priority (issue-management.md)
3. Sync issues to the project board (board-automation.md)
4. Pick the right swarm topology (mesh / hierarchical / ring / star — see swarm-topologies.md)
5. Track sprint velocity & KPIs (sprint-and-analytics.md)
6. Automate via GitHub Actions + comment commands (automation-patterns.md)
7. Verify (see below)

Verification

After running a workflow, confirm:

• Every issue acted on shows the expected label change in gh issue view <num> --json labels
• Every board-synced issue appears as a project card in the expected column
• Subtask checklists in issue bodies are valid Markdown (parseable; no broken - [ ] syntax)
• Progress comments posted by automation are idempotent — re-running doesn't duplicate
• gh api rate_limit shows headroom (running this against thousands of issues can exhaust the 5,000 req/hr limit; see troubleshooting.md)
• Closed-as-stale issues match the explicit STALE_DATE cutoff — spot-check three

References

• references/issue-management.md — creation, batch, issue→swarm, triage, decomposition, progress, stale management
• references/board-automation.md — board init, mapping config, sync, smart card management, views, dashboards
• references/sprint-and-analytics.md — what gets tracked, sprint setup, kanban/scrum, analytics, KPIs, release planning
• references/swarm-topologies.md — multi-board sync, dependencies/epics, cross-repo, team collab, topology guide
• references/templates.md — issue templates (integration / bug / feature / swarm task)
• references/automation-patterns.md — GitHub Actions workflow, specialised swarms, operational guidance, complete worked example, command quick-reference
• references/troubleshooting.md — sync, performance, recovery, GitHub constraints, required permissions, security checklist (Command Authorization, Rate Limiting, Audit Logging, Data Privacy, Access Control, Webhook Security)

Related Skills

• github-pr-workflow — link issues to pull requests automatically
• github-release-management — coordinate release issues and milestones
• sparc-orchestrator — complex project coordination workflows
• sparc-tester — automated testing workflows for issues

External Documentation

• GitHub CLI manual
• GitHub Projects (v2) docs
• ruv-swarm
• claude-flow

Maintained by Claude Code. Version 2.1.0.