name: Auto-merge field-journal PRs

on:
  pull_request:
    types: [opened, synchronize]
    paths:
      - 'skills/field-journal/**'

jobs:
  validate-and-merge:
    runs-on: ubuntu-latest
    permissions:
      contents: write
      pull-requests: write
    env:
      PR_NUMBER: ${{ github.event.pull_request.number }}
      PR_TITLE: ${{ github.event.pull_request.title }}
      PR_SUBJECT: ${{ format('[field-journal] {0}', github.event.pull_request.title) }}
    
    steps:
      - name: Checkout PR
        uses: actions/checkout@3d3c42e5aac5ba805825da76410c181273ba90b1 # v7.0.1
        with:
          fetch-depth: 0

      - name: Validate PR safety
        id: validate
        run: |
          set -e
          
          # 1. 只允许修改 field-journal 目录下的 .md 文件
          CHANGED_FILES=$(gh pr diff "$PR_NUMBER" --name-only)
          echo "Changed files:"
          echo "$CHANGED_FILES"
          
          # 严格白名单：只允许 field-journal 下的 .md 文件
          INVALID_FILES=$(echo "$CHANGED_FILES" | grep -v '^skills/field-journal/.*\.md$' || true)
          if [ -n "$INVALID_FILES" ]; then
            echo "❌ BLOCKED: PR 修改了非 field-journal 文件"
            echo "$INVALID_FILES"
            echo "valid=false" >> $GITHUB_OUTPUT
            exit 1
          fi
          
          # 不允许修改 CONTRIBUTE-BACK.md 和 _template.md（系统文件）
          SYSTEM_FILES=$(echo "$CHANGED_FILES" | grep -E '(CONTRIBUTE-BACK\.md|_template\.md)' || true)
          if [ -n "$SYSTEM_FILES" ]; then
            echo "❌ BLOCKED: 不允许修改系统文件"
            echo "$SYSTEM_FILES"
            echo "valid=false" >> $GITHUB_OUTPUT
            exit 1
          fi
          
          # 2. 文件数量限制（单次 PR 最多 5 个文件）
          FILE_COUNT=$(echo "$CHANGED_FILES" | wc -l)
          if [ "$FILE_COUNT" -gt 5 ]; then
            echo "❌ BLOCKED: 单次 PR 文件数超过 5 个（实际 $FILE_COUNT）"
            echo "valid=false" >> $GITHUB_OUTPUT
            exit 1
          fi
          
          # 3. 逐文件安全扫描
          for file in $CHANGED_FILES; do
            if [ ! -f "$file" ]; then
              continue
            fi
            
            echo "--- Scanning: $file ---"
            
            # 3a. Prompt injection 检测（宽模式）
            if grep -iE "(ignore previous|ignore all|you are now|forget (all|everything)|disregard|new instructions|system prompt|override|jailbreak|DAN mode|bypass safety)" "$file"; then
              echo "❌ BLOCKED: $file 包含 prompt injection 特征"
              echo "valid=false" >> $GITHUB_OUTPUT
              exit 1
            fi
            
            # 3b. HTML/JS 注入检测
            if grep -iE "(<script|<iframe|<object|<embed|<form|<input|javascript:|onerror=|onload=|onclick=|onmouseover=|onfocus=|data:text/html)" "$file"; then
              echo "❌ BLOCKED: $file 包含 HTML/JS 注入代码"
              echo "valid=false" >> $GITHUB_OUTPUT
              exit 1
            fi
            
            # 3c. 可执行代码/shell 命令伪装检测
            if grep -E "^(#!/|import os|import subprocess|require\('child_process|from os import|exec\(|eval\(|os\.system|subprocess\.|Runtime\.getRuntime|ProcessBuilder|powershell -e|cmd /c|curl .* \| (bash|sh)|wget .* -O- \| (bash|sh))" "$file"; then
              echo "❌ BLOCKED: $file 包含可执行代码/恶意命令"
              echo "valid=false" >> $GITHUB_OUTPUT
              exit 1
            fi
            
            # 3d. 编码绕过检测（base64 编码的命令）
            if grep -iE "(base64 -d|base64 --decode|atob\(|Buffer\.from\(.*base64|echo .* \| base64)" "$file"; then
              echo "❌ BLOCKED: $file 包含 base64 编码绕过特征"
              echo "valid=false" >> $GITHUB_OUTPUT
              exit 1
            fi
            
            # 3e. 未脱敏 token/key 检测（扩展模式）
            if grep -iE "(AKIA[0-9A-Z]{16}|sk-[a-zA-Z0-9]{20,}|ghp_[a-zA-Z0-9]{36}|gho_[a-zA-Z0-9]{36}|npm_[a-zA-Z0-9]{36}|xox[baprs]-[a-zA-Z0-9-]+|ya29\.[a-zA-Z0-9_-]+|AIza[a-zA-Z0-9_-]{35}|[0-9]+-[a-zA-Z0-9_]{32}\.apps\.googleusercontent\.com|sk_live_[a-zA-Z0-9]{24,}|rk_live_[a-zA-Z0-9]{24,}|sq0atp-[a-zA-Z0-9_-]{22})" "$file"; then
              echo "❌ BLOCKED: $file 包含疑似真实 API key/token"
              echo "valid=false" >> $GITHUB_OUTPUT
              exit 1
            fi
            
            # 3f. 真实 IP 地址检测（非 RFC1918 私有地址、非示例地址）
            # 允许: 10.x.x.x, 172.16-31.x.x, 192.168.x.x, 127.x.x.x, 0.0.0.0
            if grep -oE '\b([0-9]{1,3}\.){3}[0-9]{1,3}\b' "$file" | grep -vE '^(10\.|172\.(1[6-9]|2[0-9]|3[01])\.|192\.168\.|127\.|0\.0\.0\.0|255\.|224\.)' | grep -vE '^(198\.51\.100\.|203\.0\.113\.|192\.0\.2\.)' | head -5 > /tmp/real_ips; then
              if [ -s /tmp/real_ips ]; then
                echo "⚠️ WARNING: $file 可能包含未脱敏的真实 IP："
                cat /tmp/real_ips
                echo "❌ BLOCKED: 请将真实 IP 替换为 10.x.x.x 或 192.168.x.x"
                echo "valid=false" >> $GITHUB_OUTPUT
                exit 1
              fi
            fi
            
            # 3g. 文件大小检查（单文件 50KB 上限）
            SIZE=$(wc -c < "$file")
            if [ "$SIZE" -gt 51200 ]; then
              echo "❌ BLOCKED: $file 超过 50KB（$SIZE bytes）"
              echo "valid=false" >> $GITHUB_OUTPUT
              exit 1
            fi
            
            # 3h. Markdown 结构验证（必须以 # 开头，证明是正常 journal）
            FIRST_CONTENT_LINE=$(grep -m1 -v '^$' "$file" || echo "")
            if [[ ! "$FIRST_CONTENT_LINE" =~ ^#\  ]]; then
              echo "❌ BLOCKED: $file 不是有效的 field-journal 格式（必须以 # 标题开头）"
              echo "valid=false" >> $GITHUB_OUTPUT
              exit 1
            fi
            
            # 3i. 链接/URL 安全检测（不允许外部可执行链接）
            if grep -iE "(https?://.*\.(exe|msi|bat|cmd|ps1|sh|py|rb|pl|vbs|js)(\?|$|[\"' ]))" "$file"; then
              echo "❌ BLOCKED: $file 包含指向可执行文件的外部链接"
              echo "valid=false" >> $GITHUB_OUTPUT
              exit 1
            fi
            
            echo "✅ $file 通过安全检查"
          done
          
          echo ""
          echo "✅ 所有文件通过安全审核"
          echo "valid=true" >> $GITHUB_OUTPUT
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}

      - name: Auto-merge if valid
        if: steps.validate.outputs.valid == 'true'
        run: |
          gh pr merge "$PR_NUMBER" \
            --auto --squash \
            --subject "$PR_SUBJECT"
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}

      - name: Comment on failure
        if: steps.validate.outputs.valid == 'false'
        run: |
          gh pr comment "$PR_NUMBER" \
            --body "❌ **自动审核未通过**

          本 PR 未通过安全检查，无法自动合并。可能的原因：
          - 修改了 field-journal 以外的文件
          - 内容包含可疑的 prompt injection 特征
          - 内容包含未脱敏的 API key/token
          - 单个文件超过 50KB

          请检查并修正后重新提交。"
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}