技能 find-bugs
📦

find-bugs

安全 ⚙️ 外部命令🌐 網路存取

尋找錯誤與安全漏洞

也可從以下取得: getsentry,davila7

此技能使用系統化的多階段方法審查程式碼變更,以找出錯誤、安全漏洞和品質問題。它會繪製攻擊面、執行安全檢查清單,並提供帶有證據和修復建議的優先排序結果。

支援: Claude Codex Code(CC)
⚠️ 65
1

下載技能 ZIP

2

在 Claude 中上傳

前往 設定 → 功能 → 技能 → 上傳技能

3

開啟並開始使用

測試它

正在使用「find-bugs」。 Find bugs in the current branch changes

預期結果:

## 程式碼審查發現

### 安全漏洞

**src/auth.js:42 - SQL 注入風險**
- 嚴重性:嚴重
- 問題:使用者輸入直接串接在 SQL 查詢中
- 證據:第 42 行沒有參數化,直接使用 user.id
- 修復:使用參數化查詢或 ORM
- 參考:OWASP A03:2021

### 錯誤

**src/parser.js:15 - 空值參考**
- 嚴重性:中
- 問題:嘗試存取 undefined 的屬性
- 證據:存取 user.preferences 前沒有空值檢查
- 修復:新增空值檢查或使用可選串連

正在使用「find-bugs」。 Review for security issues only

預期結果:

## 安全稽核

已審查 12 個檔案。發現 2 個安全問題:

1. **api/handler.js:28** - 命令注入(嚴重)
2. **utils/crypto.js:8** - 弱隨機(高)

身份驗證、授權或資料驗證中未發現問題。

安全審計

安全
v1 • 2/25/2026

All static findings evaluated as false positives. External commands (git/gh CLI) are legitimate code review tooling. Network URL is documentation only. Cryptographic warnings are keyword false positives. Skill is safe for publication.

1
已掃描檔案
87
分析行數
6
發現項
1
審計總數

高風險問題 (2)

False Positive: Weak Cryptographic Algorithm (Line 3)
Static scanner flagged 'weak cryptographic algorithm' but line 3 contains only 'security vulnerabilities' in description text - no cryptographic code present
False Positive: Weak Cryptographic Algorithm (Line 77)
Static scanner flagged 'weak cryptographic algorithm' but line 77 contains only '* **File:Line** - Brief description' - no cryptographic code present
中風險問題 (2)
False Positive: External Commands (Line 23)
git diff command with $(gh repo view...) substitution is legitimate code review tooling. gh CLI command is hardcoded, not user-controlled input.
False Positive: Network Hardcoded URL (Line 4)
URL in YAML frontmatter is source reference only, not an active network request

風險因素

⚙️ 外部命令 (3)
🌐 網路存取 (1)
審計者: claude

品質評分

38
架構
100
可維護性
85
內容
23
社群
75
安全
91
規範符合性

你能建構什麼

提交前程式碼審查

在提交Pull Request之前執行,以儘早發現錯誤和安全問題

Pull Request 安全稽核

系統性審查傳入的變更,以找出安全漏洞和 OWASP 問題

自動化品質閘道

整合至 CI/CD 管線,以封鎖具有關鍵安全問題的提交

試試這些提示

基本錯誤審查
Find bugs in the current branch changes
安全重點審查
Review the changes on this branch for security vulnerabilities. Check for injection risks, authentication issues, and data exposure.
完整程式碼稽核
Perform a full security and quality audit of the changes on this branch. Map attack surfaces, run the security checklist, and provide prioritized findings with evidence and fixes.
快速品質檢查
Quickly scan the branch changes for critical bugs and high-severity security issues. Skip stylistic issues.

最佳實務

  • 在每個 Pull Request 之前執行,以儘早發現問題
  • 在合併前解決嚴重和高嚴重性的發現
  • 使用發現結果教育團隊了解安全模式
  • 整合至 CI 以防止安全回歸

避免

  • Ignoring findings without investigation because they seem false
  • Only running when explicitly asked instead of proactively
  • Treating all findings as equally urgent instead of prioritizing
  • Making changes directly instead of letting user decide

常見問題

What does this skill actually check?
It checks for bugs, security vulnerabilities like injection and XSS, authentication issues, authorization flaws, and general code quality problems using a systematic security checklist.
Does it modify my code?
No. This skill only analyzes and reports findings. It never makes changes - you decide what to fix.
What branch does it compare against?
It compares the current branch HEAD against the repository default branch (usually main or master).
Can it find all security issues?
No. It performs static analysis of diffs and cannot detect runtime issues, dependency vulnerabilities, or complex logic flaws that require execution.
Does it work with any language?
Yes. The security checklist is language-agnostic and applies general security principles to any code being reviewed.
How long does a review take?
Time varies with change size. Small PRs take seconds. Large changes with many files may take several minutes for thorough analysis.

開發者詳情

檔案結構

📄 SKILL.md