AI Decision Coach — Agent Skill

Eiliya · March 20, 2026 · 1 min read · agent-docs

AI Decision Coach — Agent Skill

This skill enables autonomous agents to perform structured decision analysis using the AI Pass API. Useful for workflow routing, intelligent orchestration, and generating recommendations for human review.

Endpoint

POST https://aipass.one/apikey/v1/chat/completions
Authorization: Bearer $AIPASS_API_KEY

Get your API key: Developer Dashboard

Quick Start

import requests, json, os

def analyze_decision(scenario: str) -> dict:
    r = requests.post(
        "https://aipass.one/apikey/v1/chat/completions",
        headers={"Authorization": f"Bearer {os.environ['AIPASS_API_KEY']}"},
        json={
            "model": "gpt-5-mini",
            "temperature": 1,
            "max_tokens": 16000,
            "messages": [
                {"role": "system", "content": "Analyze this decision. Return JSON: {recommendation, reasoning, confidence, trade_offs}"},
                {"role": "user", "content": scenario}
            ]
        }
    )
    c = r.json()["choices"][0]["message"]["content"]
    if "```" in c:
        c = c.split("```")[1].lstrip("json")
    return json.loads(c.strip())

result = analyze_decision("Should we deploy to production now or wait for more testing?")
print(result["recommendation"])  # "wait for more testing"
print(result["confidence"])      # 0.78

Supported Models

Model	Use case
`gpt-5-mini`	Standard decisions, cost-effective
`gpt-5`	Complex strategic decisions
`claude-opus-4-6`	Nuanced ethical decisions

Always use temperature: 1 and max_tokens: 16000 with GPT-5 models.

Integration Pattern

# Smart workflow router
def route_task(task: dict) -> str:
    if task["complexity"] > 7:
        decision = analyze_decision(f"How to handle: {task['description']}")
        return decision["recommendation"]
    return "auto_handle"

Related Apps

Skill File

# AI Decision Coach — Agent Skill

## Purpose
Structured decision analysis for autonomous agents. Returns JSON recommendations with confidence scores.

## API
POST https://aipass.one/apikey/v1/chat/completions
Authorization: Bearer $AIPASS_API_KEY
Get key: https://aipass.one/panel/developer.html

## Usage

```python
import requests, json, os

def analyze_decision(scenario, options=None, values=None):
    content = f'Decision: {scenario}'
    if options: content += f'\nOptions: {options}'
    if values: content += f'\nValues: {values}'
    
    r = requests.post(
        'https://aipass.one/apikey/v1/chat/completions',
        headers={'Authorization': f'Bearer {os.environ["AIPASS_API_KEY"]}'},
        json={
            'model': 'gpt-5-mini',
            'temperature': 1,
            'max_tokens': 16000,
            'messages': [
                {'role': 'system', 'content': 'Analyze and return JSON: {recommendation, reasoning, confidence, trade_offs}'},
                {'role': 'user', 'content': content}
            ]
        }
    )
    c = r.json()['choices'][0]['message']['content']
    if '```' in c: c = c.split('```')[1].lstrip('json')
    return json.loads(c.strip())

result = analyze_decision('Should we deploy now or wait for more QA?')
print(result['recommendation'])  # 'wait for more QA'
print(result['confidence'])      # 0.82
```

## Related Apps
- https://aipass.one/apps/decision-coach
- https://aipass.one/apps/conversation-coach

Download Skill File