agentleFS
khangledoanminhkhangledoanminh/agentskillslab/skills/security-auditor/SKILL.md

security-auditor

Skill1 starsChanged 3 months agoMITView on GitHub

Audit a repository for security issues: secrets, dangerous APIs, unsafe shell commands, insecure configs, and suspicious patterns, then produce an evidence-based severity-rated findings report. Use when the user asks to audit, scan, or review a repo for security vulnerabilities or risks.

  • Pipes a download into a shell

54 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use proofreader, refactoring-guide and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • xcrrrSkill

    proofreader

    Use this skill when reviewing written content for grammar, spelling, punctuation, style consistency, and tone—before publishing, submitting, or sending. Trigger phrases: 'proofread this', 'check my writing', 'review this for errors', 'edit this email/report/essay'. Do NOT use when structural rewrites or content changes are needed—proofreading fixes surface errors, not substantive problems.

    Also in 3 projects that use this

  • peopleforresterSkill

    refactoring-guide

    Safe refactoring patterns and techniques. Use when the user wants to improve code structure, reduce duplication, or make code more maintainable without changing behavior. Emphasizes incremental changes with test coverage.

    Also in 3 projects that use this

  • zwjbendnSkill

    habit-builder

    Designs and manages healthy habits for sustainable weight loss through lifestyle change. Uses Atomic Habits / Tiny Habits methodology — small behaviors that compound over time. Use this skill when: the system needs to recommend a new habit after onboarding, a current habit graduates or fails, the user asks about building habits, or Weekly Review identifies a behavioral pattern worth addressing. Also use when evaluating habit progress, deciding when a habit has "graduated," or handling a restart after the user falls off. This skill does not send its own reminders — habit check-ins are woven into meal conversations managed by Notification Composer. This skill owns the check-in logic (when, how often, tone); Notification Composer provides the conversation vehicle.

    Also in 3 projects that use this

  • juanlurgSkill

    sql-optimizer

    Analyzes SQL queries and generates an optimization report notebook with explanations, identified issues, and rewritten queries. Use when the user wants to optimize, explain, improve, or debug a SQL query, or mentions query performance, slow queries, or SQL optimization.

    Also in 3 projects that use this

  • belokonmSkill

    data-scientist

    Expert in statistical analysis, predictive modeling, machine learning, and data storytelling to drive business insights.

    Also in 3 projects that use this

  • glinckerSkill

    qa-engineer

    QA specialist agent for test planning, execution, and regression analysis

    Also in 3 projects that use this

The file

Files
SKILL.md

Security Auditor

Audit repository code với sự kết hợp giữa scanner determinist (scripts) và phân tích của AI. Scanner chạy thật, tìm pattern thật; AI xếp hạng severity, loại trừ false positive và viết báo cáo. Tuyệt đối không invent findings — nếu không quét được, trả về trạng thái not_scanned.

Quy trình (6 bước)

  1. Chuẩn bị: xác định root repo cần audit. Chạy scripts/audit.py với root dir.
  2. Review output JSON: mỗi finding có rule_id, severity, file, line, evidence, why.
  3. Phân tích từng finding (bước AI): xác nhận thật/dương tính giả, nâng/hạ severity dựa trên ngữ cảnh (VD: secret trong test fixture dummy → hạ xuống INFO).
  4. Bổ sung phân tích kiến trúc (AI, optional): coupling permission, trust boundary, nơi input không tin cậy vào hệ thống.
  5. Viết báo cáo security-report.md: bảng findings đã xác nhận + khuyến nghị fix ưu tiên theo severity.
  6. Verify không leak: báo cáo chỉ chứa evidence đã snip; không copy secret thật vào báo cáo.

Severity framework

SeverityNghĩaVí dụ
CRITICALExploitable ngay, mất credential/controlSecret thật trong source, RCE path
HIGHNguy hiểm cao, cần context để khai thácDangerous API với input user, curl|sh
MEDIUMRisk có điều kiệnUnsafe deserialization nội bộ, config yếu
LOWBest-practice violationHardcoded URL nội bộ, comment chứa key cũ
INFOĐáng chú ý, không phải lỗiDependency cũ chưa có CVE

Quy tắc phân tích AI

  • Evidence trước, kết luận sau. Mỗi finding trong báo cáo phải tham chiếu file + line.
  • Nếu scanner không chạy được (thiếu tool): ghi not_scanned cho nhóm rule đó, không giả danh sách rỗng.
  • Dương tính giả phải được giải thích ("secret là dummy trong test fixture").
  • Khuyến nghị fix cụ thể, có ví dụ code an toàn thay thế.

Scripts

  • scripts/audit.py <repo-root> [--output findings.json] — chạy tất cả nhóm rule SEC-001..006, output JSON.

Referências

  • RULES.md — chi tiết từng rule và ví dụ pattern
  • REPORT-TEMPLATE.md — template báo cáo chuẩn

Examples

python3 scripts/audit.py /path/to/repo --output findings.json
# findings.json: {"scanned": 14, "findings": [...], "not_scanned": []}

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.