feat(configs): add code_reviewer skill and coding_harness pipeline

- code_reviewer.yaml: Verifier Agent skill config for adversarial review
  with structured output schema for ReviewFeedback format
- coding_harness.yaml: Example pipeline with adversarial loop
  develop → test → review (Worker↔Verifier) → archive

configs/pipelines/coding_harness.yaml

@@ -0,0 +1,63 @@
+name: coding_harness
+version: "1.0"
+description: "Coding pipeline with adversarial review loop - Worker ↔ Verifier 对抗闭环"
+
+stages:
+  # 阶段 1：Developer 编写代码
+  - name: develop
+    agent: developer_agent
+    action: implement_feature
+    outputs:
+      - code
+      - test_files
+    timeout_seconds: 600
+    retry_count: 1
+
+  # 阶段 2：Tester 运行测试
+  - name: test
+    agent: tester_agent
+    action: run_tests
+    depends_on:
+      - develop
+    inputs:
+      code: "${develop.code}"
+      test_files: "${develop.test_files}"
+    outputs:
+      - test_results
+    timeout_seconds: 300
+    retry_count: 2
+
+  # 阶段 3：代码审查（对抗模式）
+  # Worker (developer_agent) 产出 → Verifier (code_reviewer) 审查 → 不通过则打回修复
+  - name: review
+    agent: developer_agent
+    action: fix_code_issues
+    verifier: code_reviewer
+    depends_on:
+      - test
+    max_adversarial_rounds: 3
+    feedback_mode: "structured+natural"
+    escalate_on_exhaust: human_approval
+    inputs:
+      code: "${develop.code}"
+      test_results: "${test.test_results}"
+    outputs:
+      - final_code
+      - review_report
+    timeout_seconds: 900
+
+  # 阶段 4：归档提交
+  - name: archive
+    agent: archiver_agent
+    action: commit_and_push
+    depends_on:
+      - review
+    inputs:
+      code: "${review.final_code}"
+    timeout_seconds: 120
+    continue_on_failure: false
+
+variables:
+  target_branch: main
+  require_approval: true
+  commit_message_prefix: "feat"

configs/skills/code_reviewer.yaml

@@ -0,0 +1,99 @@
+name: code_reviewer
+agent_type: dynamic_tool_chain
+version: "1.0.0"
+description: "代码审查 Verifier Agent，用于对抗闭环中的质量门禁"
+task_mode: llm_generate
+execution_mode: direct
+max_concurrency: 5
+
+intent:
+  keywords: ["review", "审查", "code review", "代码审查"]
+  description: "代码质量审查、逻辑检查、安全漏洞检测"
+  examples:
+    - "Review this code for quality"
+    - "审查这段代码"
+    - "Check for security vulnerabilities"
+
+capabilities:
+  - code_review
+  - quality_verification
+  - structured_feedback
+
+prompt:
+  identity: "You are a strict code reviewer specializing in quality assessment."
+  instructions: |
+    Review the provided code output for:
+    1. **Logic correctness** - edge cases, error handling, boundary conditions
+    2. **Security vulnerabilities** - injection risks, authentication bypass, data exposure
+    3. **Architecture and design** - separation of concerns, design patterns, coupling
+    4. **Test coverage** - are tests comprehensive, do they cover edge cases
+    5. **Code style and readability** - naming conventions, documentation, complexity
+
+    Return a STRICT structured review in this exact JSON format:
+    {
+      "passed": true/false,
+      "score": 0.0-1.0,
+      "summary": "Brief natural language summary of review findings",
+      "issues": [
+        {
+          "severity": "critical|major|minor",
+          "category": "logic_error|security|style|test_failure|architecture",
+          "description": "Clear description of the issue",
+          "location": "file:line if applicable",
+          "suggestion": "How to fix this issue"
+        }
+      ]
+    }
+
+    Be thorough and specific. If there are no issues, set passed=true and issues=[].
+
+llm:
+  model: "default"
+  temperature: 0.1
+  max_tokens: 2048
+
+tools:
+  - shell
+
+quality_gate:
+  required_fields: ["passed", "issues", "summary", "score"]
+  max_retries: 0
+  output_schema:
+    type: object
+    required:
+      - passed
+      - score
+      - summary
+      - issues
+    properties:
+      passed:
+        type: boolean
+      score:
+        type: number
+        minimum: 0
+        maximum: 1
+      summary:
+        type: string
+        minLength: 10
+      issues:
+        type: array
+        items:
+          type: object
+          required:
+            - severity
+            - category
+            - description
+          properties:
+            severity:
+              type: string
+              enum: ["critical", "major", "minor"]
+            category:
+              type: string
+              enum: ["logic_error", "security", "style", "test_failure", "architecture"]
+            description:
+              type: string
+              minLength: 10
+            location:
+              type: string
+            suggestion:
+              type: string