| name | agent-coordination |
| description | Coordinate multiple agents for software development across any language. Use for parallel execution of independent tasks, sequential chains with dependencies, swarm analysis from multiple perspectives, or iterative refinement loops. Handles Python, JavaScript, Java, Go, Rust, C#, and other languages. |
Agent Coordination
Coordinate multiple agents efficiently for complex development tasks across any programming language.
When to use
- The task has multiple workstreams (parallel, sequential, swarm, or hybrid).
- You need explicit handoff and convergence between specialized agents.
- You want throughput gains without losing validation quality.
Quick Start
Choose your coordination strategy:
Parallel - Independent tasks → See PARALLEL.md
Sequential - Dependent tasks → See SEQUENTIAL.md
Swarm - Multi-perspective analysis → See SWARM.md
Hybrid - Multi-phase workflows → See HYBRID.md
Iterative - Progressive refinement → See ITERATIVE.md
Available Agents
| Agent | Best For |
|---|
| code-reviewer | Quality assessment, standards |
| test-runner | Execute tests, verify functionality |
| feature-implementer | Build new capabilities |
| refactorer | Improve existing code |
| debugger | Diagnose and fix issues |
| security-auditor | Find vulnerabilities |
| performance-optimizer | Speed and efficiency |
| loop-agent | Orchestrate iterations |
Basic Workflow
- Plan & Issue Creation: Before starting any tasks, propose or select an issue utilizing the structured GitHub issue templates:
- For code modifications, features, or refactoring, use
🛠️ Coding Change (labels: coding, radio-play).
- For profiling, optimization, or performance checks, use
⚡ Performance Change (labels: perf, learning).
- For agent skills, plans, or harness modifications, use
🤖 Agent/Harness Change (labels: agent, harness).
- Choose strategy based on task structure
- Select agents matching required capabilities
- Execute: Implement tasks, tracking progress in
plans/GOAP_STATE.md. Verify with quality gates between phases.
- Write handoff notes to
analysis/handoffs/ with ownership and next action
- Validate outputs before proceeding
- Synthesize results into a single result
Language Support
This coordination skill works with:
- Python (Django, Flask, FastAPI)
- JavaScript/TypeScript (Node.js, React, Vue)
- Java (Spring, Jakarta EE)
- Go (Gin, Echo)
- Rust (Actix, Rocket)
- C# (.NET, ASP.NET Core)
Common Patterns
Analysis + Execution:
1. Swarm analysis (parallel agents gather insights)
2. Sequential execution (apply findings)
3. Parallel validation (verify results)
Test-Driven Workflow:
1. test-runner: Run existing tests
2. feature-implementer: Add functionality
3. test-runner: Verify implementation
4. code-reviewer: Quality check
Performance Optimization:
Loop with performance-optimizer until:
- Metrics meet targets
- No more optimizations found
- Max iterations reached
Quality Gates
Between each phase, verify:
- Proposed changes are explicitly tracked via the appropriate GitHub issue template.
- Run
bash scripts/quality_gate.sh as the default verification path.
- Code compiles/parses correctly.
- Tests pass with adequate coverage.
- At least one relevant automated test is executed and passing per coding session.
- Security scans clean.
- Performance acceptable.
- No regressions introduced.
Next Steps
Read the specific coordination pattern that matches your task structure. Each pattern includes detailed workflows, examples, and quality criteria.
For this repository, keep implementation notes synced in plans/090-recent-improvements/RECENT.md and compact outcomes in analysis/learnings/2026-04-16-real-movie-sweep.md after coordinated work lands.
