Automate post-implementation refinement to tighten test coverage and reduce complexity.
npx skills add https://github.com/z23cc/cc-code --skill cc-refinement
Please help me install this Agent Skill. Skill: cc-refinement Source: https://github.com/z23cc/cc-code/tree/main/skills/cc-refinement Command: npx skills add https://github.com/z23cc/cc-code --skill cc-refinement
Post-implementation refinement loop — quality metrics, performance budgets, edge case hardening. Use AFTER initial implementation passes TDD.
Run the refinement loop after initial implementation passes TDD to tighten quality and resilience.
High-intent search queries and answers about installing and using this skill.
Code refinement after tests pass is automated via a post-implementation refinement loop that tightens test coverage, reduces complexity, and hardens edge cases. It enforces measurable targets like ≥80% coverage and zero high vulnerabilities.
Measurable quality targets enforced during refinement include ≥80% test coverage, cyclomatic complexity (CC) ≤ 10, and zero high or critical security vulnerabilities. Performance budgets are also applied to ensure code meets resilience standards.
Yes, mypy and eslint are explicitly required to harden edge cases after TDD passes. The refinement loop uses these tools alongside pytest and bandit to measure quality dimensions and drive automated improvements.
The best way to reduce code complexity and improve security is using automated guardrails and standard checklists within a refinement loop. This workflow applies security hardening via bandit and complexity reduction targeting CC ≤ 10.
The refinement workflow supports both automated passes via /autoimmune and manual refinement via /refine workflows. Both paths enforce the same measurable targets for test coverage, complexity, and security vulnerabilities.
A refinement loop is needed after TDD passes to improve quality and performance budgets beyond basic functionality. It tightens test coverage, reduces complexity, and hardens edge cases that initial TDD implementation may leave exposed.
sizukutamagoSimplify implementation code while preserving passing test suites.
chiruu12Monitor production apps for errors and regressions after deployment.
infiquetraRun pytest, ruff, mypy, and bandit in parallel for Python projects.
ApothecaryManScan React codebases for security, performance, correctness, and architecture issues.
