coding-agent-user-rules

What This Skill Does

The coding-agent-user-rules skill acts as a foundational configuration layer for your AI development environment. It is designed to formalize the "house rules" or architectural preferences you want the OpenClaw coding agent to follow during its session. By installing this, you effectively inject a set of persistent instructions into the agent's system prompt, ensuring that it adheres to your preferred coding style, library choices, and safety guidelines without you needing to repeat these requirements in every single interaction.

Installation

To install this skill, run the following command in your OpenClaw terminal: clawhub install openclaw/skills/skills/mpaicoding-byte/coding-agent-user-rules

Use Cases

This skill is ideal for developers who want to maintain consistency across multiple coding sessions. For example, if you consistently work in a team that enforces strict type hinting in Python, or if you prefer functional programming paradigms over object-oriented ones, this skill forces the agent to adopt those standards by default. It is also useful for setting up "negative constraints," such as explicitly forbidding the agent from using specific deprecated libraries or requiring it to always include unit tests with every function generated.

Example Prompts

1. "Apply my standard user rules: always use Python 3.11 type hints and avoid external dependencies for basic string manipulation."
2. "Refactor the current module based on the coding agent user rules; ensure all error handling follows the custom logging structure defined in my settings."
3. "Can you audit the existing project structure against the coding-agent-user-rules configuration to identify where we are deviating from my style guide?"

Tips & Limitations

It is important to remember that this skill serves as an advisory layer for the agent. While it influences the agent's decision-making, it does not rewrite the underlying model architecture. Use this to define conventions, but always verify critical code logic. Currently, the skill is static and requires manual updates to the configuration file if your project preferences shift. Be aware that since this skill modifies the agent's behavior globally within the project, conflicting rules between different agents may lead to unexpected outputs if you have multiple rule-based skills installed simultaneously.