Four Principles
four-principles/
Four always-on rules for your coding agent: think before coding, keep it simple, change only what was asked, and loop until a check passes.
Four always-on engineering principles the agent applies to every task, not just the big ones.
Why use it
- The agent guesses instead of asking, and picks one reading of an ambiguous request silently.
- It ships 200 lines where 50 would do, and refactors code nobody asked it to touch.
- It says "done" without ever running the check.
What it does
- Think Before Coding: surface assumptions and trade-offs; ask instead of guessing.
- Simplicity First: the minimum that solves the problem; nothing speculative.
- Surgical Changes: every changed line traces to the request; adjacent code stays untouched.
- Goal-Driven Execution: reframe vague asks into checkable ones, then loop until the check passes.
Benefits
- Reviews get fast. Diffs stay small and every line traces to the ask, no archaeology.
- No surprise rewrites. Your untouched code stays untouched.
- "Done" actually means done. Verified, not hoped.
- Drops into any task. One file, every language, zero setup.
Install
Download this skill and copy the four-principles folder to .claude/skills/four-principles/ in your project, or to ~/.claude/skills/four-principles/ to use it in every project. Claude Code finds it on its own; you can also call it with /four-principles.
Part of the Agent Engineering bundle. To get the whole set as a Claude Code plugin: /plugin marketplace add endorphin-ai/hasbrains-agent-kit, then /plugin install agent-engineering@hasbrains-agent-kit.
Source: endorphin-ai/hasbrains-agent-kit. License: MIT.
#!/usr/bin/env bash
# Not used: a placeholder from Brain Lab. This skill ships no example script.
exit 0