Score prompt quality and detect anti-patterns for Claude Code.
npx skills add https://github.com/Kgan01/ghengis-skills --skill pql-validation
Please help me install this Agent Skill. Skill: pql-validation Source: https://github.com/Kgan01/ghengis-skills/tree/main/plugins/ghengis-skills/skills/pql-validation Command: npx skills add https://github.com/Kgan01/ghengis-skills --skill pql-validation
Validates prompts for Claude Code by detecting anti-patterns, enforcing quality principles, and suggesting safe auto-fixes before execution.
Run the validator on your prompt draft to receive a quality score and suggested fixes.
High-intent search queries and answers about installing and using this skill.
You validate prompts for Claude Code by running the validator to score quality and detect anti-patterns across task, context, format, scope, reasoning, and agentic categories before agent dispatch. It outputs a quality score and suggested fixes.
Common prompt anti-patterns in Claude Code subagent calls span task, context, format, scope, reasoning, and agentic categories. The validator detects these quality risks and provides actionable guidance to improve prompt clarity and safety.
You can automatically fix prompt quality issues before code generation by applying the validator's auto-fix suggestions. It enforces quality principles and suggests safe fixes without sharing sensitive data.
Prompt validation works seamlessly with skill-chains and scratchpads. It provides a privacy-safe scratchpad and structured outputs like quality scores, detected anti-patterns, and suggested fixes for integration.
You should run a prompt quality check on drafts before agent dispatch, code generation, or subagent calls. Running the validator ensures your prompts are free of anti-patterns and meet quality principles before execution.
leadmind-aiGenerate Claude Code subagent configuration templates with YAML frontmatter.
alexei-ledDelegate coding tasks to an OpenAI Codex agent via the Codex CLI.
richtaborValidate agent skills' structure, frontmatter, description quality, and anti-patterns, outputting a structured report.
thecloudtipsScore prompts across five dimensions and rewrite them to fix anti-patterns.
