Bỏ qua tới nội dung
TakoAPI
Quay lại kỹ năng

clawdefender

bởi nukewire

Clawdbot Tools
7.0kLượt tải do clawskills.sh công bố
29Sao
1Lượt xem
0 lượt thíchĐăng nhập để thích kỹ năng này
—Chưa có đánh giá
Đánh giá của bạn

Tóm tắt

Security scanner and input sanitizer for AI agents.

Mô tả

Kỹ năng này làm gì

Security scanner and input sanitizer for AI agents. Detects prompt injection, command injection, SSRF, credential exfiltration, and path traversal across 90+ patterns. Runs as shell scripts that pipe into any agent workflow.

It runs as a standard shell pipe with no external dependencies, so it integrates into any existing agent workflow without additional configuration.

Khi nào nên dùng

  • Auditing all installed skills for malware before running them
  • Sanitizing API responses before an agent acts on them
  • Validating a URL before fetching to prevent SSRF
  • Filtering GitHub issue content to block embedded injection attempts
  • Blocking credential theft attempts in untrusted email content

Quy trình ví dụ

ĐẦU VÀO

User asks: sanitize this API response before the agent processes it

AGENT
  1. 1

    Pipe the API response through sanitize.sh with the --json flag

  2. 2

    Scanner checks content against prompt injection, command injection, and exfiltration patterns

  3. 3

    Flagged sections are wrapped with warning markers and withheld from processing

  4. 4

    Clean content passes through for normal agent use

ĐẦU RA

Sanitized JSON ready for the agent, with suspicious sections marked and flagged to the user

Cài đặt

clawhub install nukewire/clawdefender

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

Liên kết