mirror of
https://github.com/ruvnet/ruflo.git
synced 2026-11-09 14:04:06 +09:00
rUvandClaude Opus 4.5
1b6a1971f7
- Reorganize v3/ into implementation/ subfolders: - planning/, architecture/, migration/, integration/ - security/, adrs/, research/, swarm-plans/ - Add 15-agent concurrent swarm implementation plans: - SWARM-OVERVIEW.md: Hierarchical mesh topology, 4 phases - AGENT-SPECIFICATIONS.md: 15 agent configs with capabilities - TDD-LONDON-SCHOOL-PLAN.md: Mock-first TDD methodology - BENCHMARK-OPTIMIZATION.md: Performance targets (2.49x-7.47x) - DEPLOYMENT-PLAN.md: Release strategy, CI/CD, rollback - GITHUB-ISSUE-TRACKING.md: Issue workflow with agent replies - Update .claude/settings.json with proper hooks format - Add bash aliases for claude --dangerously-skip-permissions 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Claude Flow Benchmark System v2.0
Production-ready benchmarking for Claude Flow with real command execution and authentic metrics.
Quick Start
cd benchmark
pip install -e .
# Run real benchmark
python examples/real_swarm_benchmark.py "Build REST API"
Features
- ✅ Real Execution: Actual
./claude-flowcommands via subprocess - ✅ Stream JSON: Parses
--non-interactive --output-format stream-json - ✅ Authentic Metrics: Real tokens, timing, and resource usage
- ✅ No Mocks: 100% genuine execution
Supported Commands
./claude-flow swarm "task" --non-interactive --output-format stream-json
./claude-flow hive-mind spawn "task" --non-interactive
./claude-flow sparc run code "task" --non-interactive
Usage
from swarm_benchmark import BenchmarkEngine
engine = BenchmarkEngine(use_real_executor=True)
result = await engine.run_real_benchmark("Build microservices")
print(f"Tokens: {result.metrics['total_tokens']}")
Testing
pytest tests/
python examples/verify_real_integration.py
Version: 2.0.0 | Status: Production Ready | Real: Yes | Mocks: None