Bỏ qua tới nội dung
TakoAPI
Quay lại kỹ năng

skill-bomb-dog-sniff

bởi lvcidpsyche

Coding Agents & IDEs
994Lượt tải do clawskills.sh công bố
2Sao
0Lượt xem
0 lượt thíchĐăng nhập để thích kỹ năng này
—Chưa có đánh giá
Đánh giá của bạn

Tóm tắt

Security-first skill management for OpenClaw - like a bomb-sniffing dog for skills.

Mô tả

Kỹ năng này làm gì

A static analysis scanner for OpenClaw skills that detects malicious payloads before installation. Screens for 13 threat categories including crypto stealers, reverse shells, keyloggers, and supply chain attacks using 60+ patterns with confidence scoring. Follows a quarantine-scan-install workflow so untrusted code never runs on your system.

Blocks threats at the file level before any code executes, unlike runtime monitoring tools that only detect malicious behavior after the fact.

Khi nào nên dùng

  • Scanning a third-party skill download before installing it
  • Auditing already-installed skills after a security incident
  • Running batch scans across an entire skills directory
  • Adding skill security checks to a CI/CD pipeline
  • Dry-run checking a GitHub-sourced skill without committing to install

Quy trình ví dụ

ĐẦU VÀO

User asks: safely install the 'bird' skill from ClawHub

AGENT
  1. 1

    Downloads 'bird' to a randomized quarantine directory with restricted permissions

  2. 2

    Scans all non-binary files against 60+ detection patterns across 13 threat categories

  3. 3

    Calculates a risk score using severity and confidence multipliers

  4. 4

    Compares score against the threshold (default 39) and blocks if exceeded

  5. 5

    Moves the skill from quarantine to the skills directory and cleans up if score passes

ĐẦU RA

Skill installed successfully with a scan report showing 0 critical findings and a risk score of 12/100

Cài đặt

clawhub install lvcidpsyche/skill-bomb-dog-sniff

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

Liên kết