Audit AI agent outputs and assign a 1-10 quality score with evidence.
npx skills add https://github.com/buildingopen/bouncer --skill bouncer-buildingopen
Please help me install this Agent Skill. Skill: bouncer Source: https://github.com/buildingopen/bouncer/tree/main/skill Command: npx skills add https://github.com/buildingopen/bouncer --skill bouncer-buildingopen
On-demand Gemini quality audit for ANY work: code, videos, stories, advice, research, creative output. Use when user says "audit", "bouncer", "score", "score this", "score using gemini", "check my work", "quality check", "review my changes", "rate this", "how did I do". This skill automatically scores outputs and surfaces issues to reduce risk and rework.
Run the bouncer script by piping your assistant_text and diff_text as JSON to the tool to receive a structured score and feedback.
High-intent search queries and answers about installing and using this skill.
Run an independent quality audit by piping your assistant_text and diff_text as JSON into the bouncer script. The tool gathers context, diffs, and CLAUDE.md to return a structured verdict with a numeric score and actionable guidance.
Gemini audits can review and score creative work, videos, stories, and research outputs. The audit assigns a 1-10 quality score and surfaces issues, applicable across coding, creative work, and operations tasks.
Quick audit modes provide fast feedback for checking work, while deep audit modes perform thorough verification. Both modes output a numeric score, identified issues, and actionable guidance for improving AI-generated outputs.
Use an automated quality check when you need to reduce risk and rework on AI agent outputs. It independently scores outputs, surfaces issues, and returns a structured verdict with evidence when you say "audit" or "score this".
Yes, the Gemini audit script requires the google-genai dependency to function. It uses this dependency to gather context and diffs, returning a structured quality verdict with evidence for your AI-generated work.
allenhutchisonBuild applications with Gemini models across Python, JavaScript, Java, and Go SDKs.
AcidicSoilApply prioritized QA fixes and update the task record.
shuff57Automate coding tasks, code review, and refactoring with Claude Code CLI.
maxorericEvaluate AI-generated outputs against predefined criteria for scoring and iteration decisions.
