AI Prompt Optimizer Agent Skill — AI Pass API

Eiliya · March 22, 2026 · 2 min read · agent-docs

Overview

What this skill does

The AI Pass "AI Prompt Optimizer" skill analyzes, rewrites, and improves AI prompts to produce clearer, more specific, and more consistent outputs.
It focuses on maximizing prompt clarity, reducing ambiguity, enforcing constraints, and improving instruction structure so downstream models provide higher-quality results.

Use cases for AI agents

Automatically rewrite user-provided prompts before forwarding them to a model to increase reliability and reduce follow-up clarifications.
Provide multiple optimized prompt variants (concise, detailed, constrained) so an agent can choose the best one for a task.
Normalize prompt style and include required context, system instructions, or token/length constraints for pipelines that call multiple models.
Convert informal user requests into production-ready prompts for content generation, data extraction, summarization, or code generation.

Skill File (YAML)

name: prompt-optimizer
description: Rewrites and improves AI prompts to get better, more consistent results
version: "1.0"
auth:
  type: bearer
  token_env: AIPASS_API_KEY
endpoints:
  - name: generate
    method: POST
    url: https://aipass.one/apikey/v1/chat/completions
    headers:
      Authorization: "Bearer $AIPASS_API_KEY"
      Content-Type: application/json

Example Usage

curl example

curl -s -X POST "https://aipass.one/apikey/v1/chat/completions" \
  -H "Authorization: Bearer $AIPASS_API_KEY" \
  -H "Content-Type: application/json" \
  -H "X-Client-ID: YOUR_CLIENT_ID" \
  -d '{
    "model": "aipass-4o-chat",
    "messages": [
      {"role": "system", "content": "You are an expert prompt engineer. Rewrite the user prompt to maximize clarity, specificity, and expected output quality. Provide the optimized prompt and a 1-2 sentence rationale."},
      {"role": "user", "content": "Write me a blog post about remote work that converts readers into subscribers. Product: weekly newsletter about remote work tips."}
    ],
    "temperature": 0.2,
    "max_tokens": 400
  }'

Python example

import os
import requests

API_URL = "https://aipass.one/apikey/v1/chat/completions"
API_KEY = os.environ.get("AIPASS_API_KEY")  # ensure this is set in your environment
HEADERS = {
    "Authorization": f"Bearer {API_KEY}",
    "Content-Type": "application/json",
    "X-Client-ID": "YOUR_CLIENT_ID"
}

payload = {
    "model": "aipass-4o-chat",
    "messages": [
        {"role": "system", "content": "You are an expert prompt engineer. Rewrite the user prompt to maximize clarity, specificity, and expected output quality. Provide the optimized prompt and a 1-2 sentence rationale."},
        {"role": "user", "content": "Draft an about page for my startup focusing on sustainability and growth."}
    ],
    "temperature": 0.2,
    "max_tokens": 400
}

resp = requests.post(API_URL, headers=HEADERS, json=payload)
resp.raise_for_status()
print(resp.json())

Notes

Keep temperature low (0.0–0.3) when you want consistent, deterministic prompt rewrites.
Request multiple variants by asking the model to return numbered options if you need A/B testing.

Available Models

aipass-3.5-chat — cost-effective, fast inference for standard prompt optimization tasks.
aipass-4o-chat — balanced, high-quality chat model suitable for most prompt engineering needs.
aipass-4-creative — higher-capacity model tuned for creative rewrites and tone/style variations.
aipass-instruct-1 — instruction-tuned model focused on precise, constraint-following rewrites.

(Choose the model that matches your latency, cost, and creativity requirements.)

Get Your API Key

Sign in to the developer dashboard and create an API key: https://aipass.one/panel/developer.html → API Keys section
Always store keys securely and never embed real keys directly in code—use environment variables (AIPASS_API_KEY).

Related Apps

Prompt Optimizer app: https://aipass.one/apps/prompt-optimizer

Note: Never include real API keys in code samples. Use the placeholder $AIPASS_API_KEY for all examples and YOUR_CLIENT_ID for any client ID references.

Skill File

id: ai_pass_prompt_optimizer
name: AI Pass Prompt Optimizer
version: 1.0.0
author: acme.ai
description: Rewrites and improves AI prompts to get better, more consistent results
auth:
  type: apiKey
  in: header
  name: Authorization
  env: AIPASS_API_KEY
  prefix: "Bearer"
inputs:
  prompt:
    type: string
    description: The original prompt to rewrite and optimize
    required: true
  target_style:
    type: string
    description: Optional. Desired tone/format for the optimized prompt (e.g., concise, detailed, step-by-step)
    required: false
  model:
    type: string
    description: Optional. Model to use for the rewrite (defaults to a recommended model for AI Pass)
    required: false
    default: "gpt-4o"
  temperature:
    type: number
    description: Optional. Creativity (0.0-1.2). Lower values = more deterministic.
    required: false
    default: 0.2
  max_tokens:
    type: integer
    description: Optional. Maximum tokens for the response.
    required: false
    default: 800
functions:
  - name: optimize_prompt
    description: Rewrites and improves AI prompts to get better, more consistent results
    inputs:
      - prompt
      - target_style
      - model
      - temperature
      - max_tokens
    run:
      http:
        method: POST
        url: https://aipass.one/apikey/v1/chat/completions
        headers:
          Content-Type: application/json
          Authorization: "Bearer $AIPASS_API_KEY"
        body:
          model: "{{model}}"
          temperature: "{{temperature}}"
          max_tokens: "{{max_tokens}}"
          messages:
            - role: system
              content: >-
                You are an expert prompt engineer. Rewrite and improve the user's prompt so downstream AI models produce more accurate,
                consistent, and reliable outputs. Preserve the original intent and any constraints. Output only the final optimized prompt
                text, with no explanations or extra commentary.
            - role: user
              content: >-
                Original prompt:
                {{prompt}}
                {{#if target_style}}
                Desired style: {{target_style}}
                {{/if}}
        response:
          status_codes: [200]
          content_type: application/json
          output_path: choices[0].message.content
outputs:
  optimized_prompt:
    type: string
    description: The rewritten, optimized prompt returned by AI Pass
    from: functions.optimize_prompt.response.output_path
metadata:
  ready_for: OpenClaw
  notes: >
    Uses the $AIPASS_API_KEY environment variable for authentication. The function posts to
    https://aipass.one/apikey/v1/chat/completions using the Chat Completions schema and returns the optimized prompt.

Download Skill File