Skip to content

Comments

Add Claude Flow V3 configuration and behavioral guidelines#1

Open
proffesor-for-testing wants to merge 3 commits intomainfrom
claude/init-claude-flow-TVgeF
Open

Add Claude Flow V3 configuration and behavioral guidelines#1
proffesor-for-testing wants to merge 3 commits intomainfrom
claude/init-claude-flow-TVgeF

Conversation

@proffesor-for-testing
Copy link
Owner

Summary

This PR introduces a comprehensive configuration document (CLAUDE.md) that establishes coding standards, architectural patterns, and operational guidelines for the Claude Flow V3 system. This serves as the authoritative reference for development practices, agent orchestration, and project organization.

Key Changes

  • Behavioral Rules: Established core principles including file creation constraints, secret management, and operational discipline
  • File Organization: Defined directory structure (/src, /tests, /docs, /config, /scripts, /examples) with clear usage guidelines
  • Project Architecture: Specified Domain-Driven Design patterns, file size limits (500 lines), typed interfaces, TDD London School methodology, and event sourcing requirements
  • Build & Test Standards: Documented npm scripts and mandatory verification steps (tests, linting, build verification)
  • Security Rules: Established hardcoded secret prevention, input validation, path sanitization, and security scanning requirements
  • Concurrency Model: Defined "1 MESSAGE = ALL RELATED OPERATIONS" principle with batching requirements for todos, agents, file operations, and bash commands
  • Swarm Orchestration: Documented 3-tier model routing (Agent Booster/Haiku/Sonnet-Opus), hierarchical topology configuration, and anti-drift measures
  • CLI Reference: Provided comprehensive command reference for V3 CLI tools including agent spawning, swarm management, memory operations, and task lifecycle
  • Agent Types: Listed 60+ available agent types across core development, specialized, coordination, and SPARC methodology categories
  • Memory System: Documented AgentDB memory commands with HNSW search capabilities

Notable Implementation Details

  • Established clear separation between Claude Code's Task tool (execution) and CLI tools (coordination)
  • Defined maxAgents at 6-8 for tight coordination with hierarchical topology
  • Specified run_in_background: true for all agent Task calls to enable parallel execution
  • Documented 3-tier model routing to optimize latency and cost based on task complexity
  • Included quick setup instructions and support references for the Claude Flow ecosystem

https://claude.ai/code/session_01QurwDjgvwVJJ4oeZdXPDXN

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants