Open Government, Taiwan

```json
{
  "service_type": "platform",
  "base_url": "https://data.gov.tw",
  "auth_method": "none",
  "auth_config": {},
  "endpoints": [],
  "pricing_model": {
    "type": "free",
    "details": {}
  },
  "rate_limits": {},
  "capabilities": [
    "Provides open government data from Taiwan",
    "Data includes various categories such as AI and general governmental data",
    "Serves as a central repository for government datasets",
    "Supports data transparency and public access to government information",
    "Facilitates the reuse of government data for innovation and research"
  ],
  "raw_analysis": "Open Government, Taiwan (data.gov.tw) is a platform providing open data from the Taiwan government. It is designed for researchers, developers, policy makers, and citizens interested in accessing a wide range of governmental data categories, including AI and data. The platform is mature, serving as a comprehensive repository for Taiwan's open data initiatives. There appears to be no specific REST API detailed, and the platform is freely accessible, promoting transparency and public engagement with government data. No explicit rate limits are mentioned, highlighting its focus on being a readily accessible resource. There are no direct noted integrations with other platforms or tools, but its open data can be used in various applications."
}
```

1/3 tests passed

```json
{
  "overall": 52,
  "dimensions": {
    "token_efficiency": 7.0,
    "first_try_success": 5.0,
    "response_parseability": 4.0,
    "error_clarity": 4.0,
    "doc_quality": 5.0,
    "auth_simplicity": 8.0,
    "latency": 6.0,
    "consistency": 5.0
  },
  "pricing_normalized": {
    "score": 10.0,
    "model": "free",
    "note": "Completely free government open data platform, no registration barriers"
  },
  "issues": [
    "No robots.txt found (404) — impedes automated discovery and agent crawling",
    "No llms.txt found (404) — no LLM-optimized documentation for agent consumption",
    "Response previews show HTML error states with '[object Object]' serialization failures in error attributes",
    "Primary language is Traditional Chinese (zh-Hant-TW) — adds friction for non-Chinese-speaking agents/users",
    "No structured API evidence in checks — appears to be primarily web/HTML-based data access",
    "Meta charset typo ('charse' instead of 'charset') in HTML — indicates quality control issues",
    "No clear dataset export formats or API endpoints visible from checks"
  ],
  "recommendations": [
    "Implement robots.txt with clear crawl guidance for agents and search engines",
    "Add llms.txt with structured description of available datasets, API endpoints, and access patterns",
    "Provide English-language documentation or dataset metadata to expand agent usability",
    "Expose RESTful API or GraphQL endpoint with JSON responses instead of relying on HTML scraping",
    "Add structured dataset metadata (DCAT, Schema.org) to HTML pages for machine parseability",
    "Fix HTML validation errors (charset typo, error attribute serialization)",
    "Implement dataset-specific permalinks with JSON export options",
    "Add rate limiting documentation and API key mechanism if usage scales"
  ]
}
```