Provide behavioral guidelines for LLMs to minimize coding errors.
npx skills add https://github.com/P4rzi/PROT-RISK --skill coding-guidelines-p4rzi
Please help me install this Agent Skill. Skill: coding-guidelines Source: https://github.com/P4rzi/PROT-RISK/tree/main/.github/skills/coding-guidelines Command: npx skills add https://github.com/P4rzi/PROT-RISK --skill coding-guidelines-p4rzi
This Skill provides a set of behavioral guidelines to help reduce common coding mistakes made by LLMs, promoting more robust and efficient code development.
Follow these guidelines when writing, modifying, or reviewing code for implementation tasks.
High-intent search queries and answers about installing and using this skill.
To prevent LLM coding mistakes during code review and refactoring, apply behavioral guidelines that enforce explicit assumption stating, simplicity, and surgical changes to minimize errors. This approach ensures LLMs modify only required code lines and match existing styles.
Best practices for LLM-assisted code modification include enforcing goal-driven execution, defining clear success criteria, and cleaning up only the introduced mess. This keeps modifications minimal, necessary, and focused on the required parts of the codebase.
To enforce surgical code changes during LLM bug fixing, apply guidelines that promote modifying only the required parts of the codebase and cleaning up solely the introduced mess. This prevents over-engineering and unintended modifications to unrelated code.
You should state explicit assumptions before generating code to clarify ambiguities and establish a pre-coding thought process. This minimizes coding errors by ensuring the LLM understands the exact requirements and executes the implementation task with clear success criteria.
This approach works for general code writing, modification, and review tasks, not just specific refactoring. The guidelines apply universally to implementation and bug fixing tasks by promoting simplicity, goal-driven execution, and minimal necessary code without over-engineering.
ManancodeGuide coding tasks with explicit assumptions and verifiable success criteria.
lagz0neEnforce fresh verification evidence before claiming work completion.
maxenkoRun cargo fmt, clippy, tests, and doc checks before committing Rust changes.
carlopezzutoApply structured prompt engineering patterns to improve LLM reliability and controllability.
