coding-practices
Coding Practices
Practices specific to writing and editing code. Use alongside agentic-standards, which provides the general behavioral baseline for all interactions.
Quick reference
| Domain | Key rules | Reference |
|---|---|---|
| Coding behavior | Scope discipline, simplest approach first, read before write, verify before done, incremental development | coding-behavior.md |
| Tools | Prefer dedicated tools over shell equivalents. Parallelize independent calls. Match-based edits. | tool-preferences.md |
Core principles
- Try the simplest approach first. Complexity is a cost. Justify it.
- Read before write. Never edit code you haven't seen this session.
- Verify with evidence. Run the test. Read the output. Report what happened.
- One change at a time. Make a logical change, validate, then continue.
- No scope creep. Do what was asked. Do not add features, refactors, or improvements beyond the request.
Load the reference files above for detailed guidance.
Verification helpers
Pair with Superpowers verification-before-completion for the “no completion claims without fresh evidence” rule. This repo keeps optional automation and a report shape that works without that plugin:
- run-checks.sh — Auto-detects test runner, linter, type checker, and build for common stacks; prints a structured PASS/FAIL report (
./run-checks.sh,--tests-only,--lint-only). - Verification report template — Evidence-style checklist and adversarial probe ideas for manual or scripted runs.
Deep references
For system-level design patterns, load when building or reviewing agent tool infrastructure:
- Tool design checklist — Fail-closed tool contract, safety declarations, progressive tool discovery, and the 9-step tool execution pipeline.
More from beltonk/claude-code-agent-skills
managing-memories
Covers the full memory lifecycle — when to save, what format to use, how to organize and deduplicate, how to recall relevant memories, and what to never persist. Use at natural breakpoints to capture user preferences, corrections, and project conventions, and at session start to load relevant context.
4scaffolding-projects
Provides a structured approach to starting new features or projects. Guides the agent through understanding requirements, exploring existing code, planning, incremental implementation, and verification. Use when asked to build something new — a feature, module, service, or project — to avoid jumping into code without context.
4handing-off-sessions
Captures structured session state for resuming work in a new session or handing off to another agent. Use at the end of a session, before context limits, or when the user asks to save progress. Not needed for trivial sessions (quick questions, one-line answers).
4compacting-context
Provides a structured 9-section summarization template for compressing long conversations while preserving critical details. Use when a session approaches context limits and history must be compressed without losing user intent, file changes, errors, or next steps.
4agentic-standards
Foundational behavioral standards for any AI agent — safety/reversibility framework, output quality, memory conventions, and prompt injection defense. Applies to all agent interactions including chat, analysis, writing, debugging, and coding. Use when setting up an agent, onboarding to a new project, or when behavioral baseline guidance is needed. For coding-specific rules, also load coding-practices.
4receiving-code-review
Enforces rigorous handling of code review feedback. Classifies comments by severity, verifies claims independently, prevents blind implementation of incorrect suggestions, and handles contradictory reviews. Use when receiving review comments on a pull request or code change.
1