Summary

This repository addresses the critical issue of rapidly accumulating technical debt in LLM-assisted code development, stemming from LLMs' inherent limitations in understanding context, institutional memory, and implicit requirements. It offers a structured workflow and set of skills for Claude Code designed to mitigate these long-term code quality challenges. The solution targets engineers seeking sustainable, high-quality AI-assisted development by enforcing rigorous planning, context management, and multi-stage review cycles before code execution.

How It Works

The workflow is built upon four core principles: Context Hygiene ensures tasks receive precisely the needed information, utilizing CLAUDE.md and README.md files for indexing and decision logging, respectively, while scrubbing code comments for temporal contamination. Planning Before Execution mandates that LLM outputs are first planned, with rationale, tradeoffs, and risks documented in files to surface ambiguities early. Review Cycles involve automated checks by technical writer and quality reviewer agents to ensure plan clarity and completeness before execution. Cost-Effective Delegation uses an orchestrator to assign tasks to appropriately tiered agents (e.g., Haiku, Sonnet), reserving expensive models for complex ambiguities. The user workflow follows explore → plan → execute, leveraging skills like analyze and planner, followed by a /clear context command before execution delegates to sub-agents.

Quick Start & Requirements

Installation involves cloning the repository into the Claude Code configuration directory.

• Per-project: git clone https://github.com/solatis/claude-config .claude
• Global (new setup): git clone https://github.com/solatis/claude-config ~/.claude
• Global (existing ~/.claude): cd ~/.claude; git remote add workflow https://github.com/solatis/claude-config; git fetch workflow; git merge workflow/main --allow-unrelated-histories Requires a functional Claude Code environment. Specific LLM tiers (Haiku, Sonnet) are utilized for delegation. No other non-default prerequisites are specified.

Highlighted Details

• Explicitly designed to combat LLM-induced technical debt.
• Enforces a strict plan-then-execute paradigm with automated review gates.
• Implements tiered LLM delegation for optimized cost and performance.
• Maintains persistent rationale and decision context in file artifacts.

Maintenance & Community

No specific details regarding maintainers, community channels (Discord/Slack), sponsorships, or roadmap are present in the provided README.

Licensing & Compatibility

The repository's license is not specified in the README. Consequently, compatibility for commercial use or closed-source linking cannot be determined.

Limitations & Caveats

This workflow is highly opinionated and primarily tested for backend engineering tasks; its applicability to frontend development is unverified. Less experienced software engineers may find the structured approach introduces overhead rather than assistance. The system's functionality is dependent on the Claude Code environment and its associated agent skills.