Bỏ qua tới nội dung
TakoAPI
Quay lại kỹ năng

arc-skill-scanner

bởi trypto1019

Coding Agents & IDEs
612Lượt 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

Scan OpenClaw skills for security vulnerabilities before installing them.

Mô tả

Kỹ năng này làm gì

Scans OpenClaw skill packages for security vulnerabilities before installation. Detects credential stealers, obfuscated code, data exfiltration, prompt injection in SKILL.md files, and binary tampering. Roughly 22-26% of ClawHub skills have been flagged as containing vulnerabilities.

Manual code review of every third-party skill is impractical; this automates detection of known attack patterns across SKILL.md content, scripts, metadata, and binaries in one pass.

Khi nào nên dùng

  • Scanning a ClawHub skill before installing it
  • Auditing all currently installed skills at once
  • Verifying binary checksums after a skill updates
  • Detecting typosquatted skill names mimicking popular ones
  • Generating checksum manifests for trusted skill versions

Quy trình ví dụ

ĐẦU VÀO

User asks: scan the skill at ~/.openclaw/skills/some-skill/ before installing

AGENT
  1. 1

    Run scanner.py with --path pointing to the skill directory

  2. 2

    Analyze SKILL.md for prompt injection patterns and suspicious URLs

  3. 3

    Check scripts for network calls, env variable access, and obfuscated strings

  4. 4

    Run typosquatting check against known popular skill names

  5. 5

    Output risk level with specific findings and file locations

ĐẦU RA

Risk assessment showing CLEAN/LOW/MEDIUM/HIGH/CRITICAL rating with a list of flagged patterns and the files they were found in

Cài đặt

clawhub install trypto1019/arc-skill-scanner

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

Liên kết