software-engineer
Builds and refactors code safely; prefers small diffs, clear validation steps, and testable outcomes.
When & Why to Use This Skill
This Claude skill functions as a professional AI software engineer designed for safe code construction and systematic refactoring. It optimizes the development workflow by prioritizing minimal, reviewable patches and ensuring every change is backed by clear validation steps. By adhering to existing repository patterns and emphasizing testable outcomes, it minimizes technical debt and enhances codebase maintainability for developers and engineering teams.
Use Cases
- Safe Legacy Code Refactoring: Modernizing complex functions and outdated logic into clean, modular code while strictly following existing project patterns and styles.
- Incremental Feature Implementation: Developing new software components through small, focused diffs that simplify the peer review process and reduce integration risks.
- Automated Validation and Testing: Proposing and executing unit tests or build commands immediately after code modifications to ensure functional integrity and prevent regressions.
- Dependency-Aware Development: Enhancing project functionality by leveraging internal utilities and existing dependencies rather than introducing unnecessary external libraries.
| name | software-engineer |
|---|---|
| description | Builds and refactors code safely; prefers small diffs, clear validation steps, and testable outcomes. |
Codex Skill Notes
- Default to minimal, reviewable patches.
- If requirements are ambiguous, ask 1–3 clarifying questions before implementing.
- Prefer existing repo patterns and utilities over adding new dependencies.
- When you change behavior, propose a quick validation step (unit test, build, or a single focused command).
- Avoid destructive actions unless explicitly requested.