AI Agent Performance Auditor | NIST & ISO Aligned Audit Tool

Audit, diagnose, and optimize any AI agent. This premium .md workspace skill evaluates token efficiency, accuracy, bias, and health using NIST, ISO, and HELM standards. An AI Agent Performance Audit is a structured, multidimensional evaluation of an AI agent's current state. It measures operational token efficiency, prompt-response relevance, factual accuracy, bias exposure, source quality, and system configuration to identify critical failures and remediation priorities. How to Use This Skill

Option A: Running in OpenClaw

1. Download and unzip your purchase files.

2. Move the `agent-performance-auditor` directory directly into your local OpenClaw skills folder:

   `~/.openclaw/skills/agent-performance-auditor/`

3. Restart your OpenClaw gateway:

   `openclaw gateway restart`

4. Trigger the auditor in your active chat window:

   "Audit my customer support agent's performance."

Option B: Manual Execution (ChatGPT / Claude / Custom LLMs)

1. Open your AI platform of choice (e.g., Claude 3.5 Sonnet or GPT-4o).

2. Open the downloaded `SKILL.md` in any text editor.

3. Copy the entire contents of the file and paste it into the chat window as your system instructions.

4. Input your target agent's system prompt, 3–5 sample chat logs, or a direct session transcript.

5. The model will process the instructions and deliver your formatted Executive Summary and Technical Diagnostic Report.