Bỏ qua tới nội dung
TakoAPI
Quay lại kỹ năng

agent-skills-tools

bởi rongself

PDF & Documents
1.0kLượt tải do clawskills.sh công bố
0Lượt xem
0 lượt thíchĐăng nhập để thích kỹ năng này
—Chưa có đánh giá
Đánh giá của bạn

Tóm tắt

Security audit and validation tools for the Agent Skills ecosystem.

Mô tả

Kỹ năng này làm gì

Security audit and validation tools for the Agent Skills ecosystem. Scans skill packages for hardcoded credentials, dangerous file access patterns, external network requests, and Git history secrets. Built after a credential stealer was found in a community skill package.

Unlike generic static analysis tools, this is tailored to the Agent Skills package format and catches the specific attack patterns agents are most vulnerable to due to their trusting execution model.

Khi nào nên dùng

  • Auditing a third-party skill for hardcoded API keys before installation
  • Scanning a skill's Git history for accidentally committed secrets
  • Checking file permissions on credential files in a skill package
  • Detecting dangerous access patterns targeting ~/.ssh or ~/.aws directories
  • Validating that a skill reads credentials from environment variables instead of source files

Quy trình ví dụ

ĐẦU VÀO

User asks: Check this skill package for security issues before I install it

AGENT
  1. 1

    Run skill-security-audit.sh against the target skill directory

  2. 2

    Scan source files for hardcoded API keys, passwords, and tokens

  3. 3

    Check for file access patterns targeting ~/.ssh, ~/.aws, or ~/.config

  4. 4

    Inspect Git history for previously committed secrets

  5. 5

    Report pass/fail status for each check category

ĐẦU RA

A security audit report listing results for each check, with details on any detected vulnerabilities or confirmation that none were found

Cài đặt

clawhub install rongself/agent-skills-tools

Chạy lệnh CLI, hoặc dán prompt vào Claude Code, Cursor, Codex hay coding agent bạn đang dùng.

Liên kết