frontend-inferencer（前端推理器）

Frontend Inferencer Skill

Purpose

Infer frontend technologies by analyzing collected signals from JavaScript, DOM, HTML, and CSS analysis phases.

Input

Raw signals from Phase 2:

• javascript_signals - Global variables, DOM attributes, bundle patterns
• html_signals - Meta tags, script URLs, CSS classes
• http_signals - Headers (for SPA indicators)

Technology Categories

JavaScript Frameworks

Meta-Frameworks

CSS Frameworks

UI Component Libraries

State Management

Build Tools (from bundle patterns)

Inference Logic

def infer_frontend_technologies(signals):
    results = []

    # JavaScript Framework Detection
    for framework in FRAMEWORK_PATTERNS:
        score = 0
        evidence = []

        # Check global variables
        for global_var in framework.globals:
            if global_var in signals.javascript_signals.globals:
                score += framework.global_weight
                evidence.append(f"{global_var} detected")

        # Check DOM attributes
        for attr in framework.dom_attributes:
            if attr in signals.html_signals.dom_attributes:
                score += framework.dom_weight
                evidence.append(f"{attr} attribute found")

        # Check bundle patterns
        for pattern in framework.bundle_patterns:
            if pattern in signals.javascript_signals.script_urls:
                score += framework.bundle_weight
                evidence.append(f"Bundle pattern: {pattern}")

        if score > 0:
            results.append({
                "name": framework.name,
                "category": framework.category,
                "signals": evidence,
                "total_weight": score,
                "implies": framework.implies
            })

    # CSS Framework Detection
    for css_framework in CSS_FRAMEWORK_PATTERNS:
        matches = count_class_matches(
            signals.html_signals.css_classes,
            css_framework.patterns
        )

        if matches >= css_framework.min_matches:
            score = css_framework.base_weight + (matches * 2)
            results.append({
                "name": css_framework.name,
                "category": "CSS Framework",
                "signals": [f"{matches} class patterns matched"],
                "total_weight": score
            })

    return results

Output

{
  "skill": "frontend_inferencer",
  "results": {
    "technologies": [
      {
        "name": "React",
        "category": "JavaScript Framework",
        "version": "18.x (if detectable)",
        "signals": [
          {
            "type": "javascript_global",
            "value": "window.React detected",
            "source": "https://example.com",
            "weight": 30
          },
          {
            "type": "dom_attribute",
            "value": "data-reactroot found",
            "source": "https://example.com",
            "weight": 25
          }
        ],
        "total_weight": 55,
        "implies": []
      },
      {
        "name": "Next.js",
        "category": "Meta-Framework",
        "version": "13.x",
        "signals": [
          {
            "type": "javascript_global",
            "value": "__NEXT_DATA__ found",
            "source": "https://example.com",
            "weight": 35
          },
          {
            "type": "bundle_pattern",
            "value": "/_next/static/ paths",
            "source": "https://example.com",
            "weight": 30
          }
        ],
        "total_weight": 65,
        "implies": ["React", "Node.js"]
      },
      {
        "name": "Tailwind CSS",
        "category": "CSS Framework",
        "signals": [
          {
            "type": "css_classes",
            "value": "15 Tailwind class patterns matched (bg-*, text-*, flex, p-*, etc.)",
            "source": "https://example.com",
            "weight": 30
          }
        ],
        "total_weight": 30
      }
    ],
    "implied_technologies": [
      {
        "name": "Node.js",
        "implied_by": ["Next.js"],
        "confidence": "High"
      }
    ],
    "summary": {
      "primary_framework": "Next.js",
      "ui_library": "Tailwind CSS",
      "state_management": null,
      "build_tool": "Webpack (via Next.js)"
    }
  }
}

Version Detection

React

// From __REACT_DEVTOOLS_GLOBAL_HOOK__
window.__REACT_DEVTOOLS_GLOBAL_HOOK__.renderers.get(1).version

// From React DevTools
window.React.version

Angular

<!-- From ng-version attribute -->
<app-root ng-version="16.2.0">

Next.js

// From __NEXT_DATA__
JSON.parse(document.getElementById('__NEXT_DATA__').textContent).nextVersion

Vue.js

// From Vue instance
window.Vue.version

Confidence Calculation

Frontend Technology Confidence:

High (80-100%):
  - Direct framework global detected
  - Multiple supporting signals
  - Version information available

Medium (50-79%):
  - Bundle patterns match
  - DOM attributes present
  - No direct confirmation

Low (20-49%):
  - Only CSS class patterns
  - Single indirect signal
  - Job posting mention only

Error Handling

• Missing signals: Return empty results for category
• Conflicting signals: Include both with notes
• Version detection failure: Omit version, continue