analyzer
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
评估生成的 security-review-skill 的检测效果。
Claude Code
$ npx -y skills add xwtro0tk1t-cloud/harness --agent claude-code
Cursor
Cursor also loads subagents from .claude/agents, so copy this file there to use it in Cursor.
评估生成的 security-review-skill 的检测效果。
用已知漏洞的测试代码运行 skill,对比检测结果和 ground truth。
| 指标 | 计算 | 说明 | |-----|------|------| | **检出率 (Recall)** | TP / (TP + FN) | 真实漏洞被检出的比例 | | **精确率 (Precision)** | TP / (TP + FP) | 报告中真实漏洞的比例 | | **定位准确** | - | 行号是否准确(允许 ±3 行) |
| 分类 | 说明 | |-----|------| | **True Positive (TP)** | 正确检出的真实漏洞 | | **False Positive (FP)** | 误报(安全代码被标记为漏洞) | | **False Negative (FN)** | 漏检(真实漏洞未被检出) |
1. **准备测试用例**
{
"test_file": "test_vuln.py",
"vulnerabilities": [
{"type": "sql_injection", "line": 15}
],
"safe_code": [
{"line": 30, "reason": "使用参数化查询"}
]
}2. **运行 skill**
让 skill 审计 test_vuln.py
3. **对比结果**
4. **输出评估报告**
{
"recall": 0.9,
"precision": 0.8,
"improvements": [
"漏检 IDOR,需补充越权检测规则",
"Jinja2 误报 XSS,需加框架保护识别"
]
}根据评估结果:
| 问题 | 改进方向 | |-----|---------| | **漏检 (FN 高)** | 补充检测模式,加强危险函数识别 | | **误报 (FP 高)** | 添加框架保护识别,加强上下文分析 | | **定位不准** | 改进代码模式描述 |
Harness is an AI Agent development scaffold Meta-Skill that establishes four layers of enhancement for any project in one command: knowledge management, architecture constraints, feedback loops, and entropy management.
Analyze blind comparison results to understand WHY the winner won and generate improvement suggestions.
Compare two outputs WITHOUT knowing which skill produced them.
Evaluate expectations against an execution transcript and outputs.
Writes and runs codemod scripts that replace hardcoded visual values with token references in ui/src/index.css. Use for Phase 2…
Complete migration plan for converting command-based system to intelligent agent-based system
Use this agent to verify that tasks marked as 'review' have been properly implemented according to their specifications. This…