Bỏ qua tới nội dung
TakoAPI
Quay lại kỹ năng

skill-trust-auditor

bởi jonathanjing

Web & Frontend Development
357Lượt 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

Audit a ClawHub skill for security risks BEFORE installation.

Mô tả

Kỹ năng này làm gì

Analyzes a ClawHub skill's scripts, permissions, and metadata for security risks before you install it. Produces a trust score from 0 to 100, flags specific risky patterns with file locations, and gives a verdict and recommendation. Works with skill slugs or full ClawHub URLs.

Rather than manually reading every script in a skill repository, it automates pattern matching against a known risk reference and gives a scored verdict in one command.

Khi nào nên dùng

  • Checking a community skill before installing it on a production machine
  • Reviewing a colleague's shared skill before adding it to a shared workspace
  • Auditing any skill that requests unusual file or network access
  • Setting up an alias so every install triggers an automatic pre-check
  • Investigating a flagged skill after a security incident report

Quy trình ví dụ

ĐẦU VÀO

User asks: "Is steipete/clawhub safe to install?"

AGENT
  1. 1

    Run bash scripts/audit.sh steipete/clawhub against the target skill

  2. 2

    Scan all scripts and SKILL.md for HIGH, MEDIUM, and LOW risk patterns

  3. 3

    Calculate a trust score by applying per-pattern penalties

  4. 4

    Check whether the skill author is verified on ClawHub

  5. 5

    Format and return a structured JSON report with score, verdict, risks, and recommendation

ĐẦU RA

Trust score of 72/100, verdict INSTALL WITH CAUTION, with a specific warning about an external curl call in scripts/sync.sh line 14 and a suggested inspection command.

Cài đặt

clawhub install jonathanjing/skill-trust-auditor

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

Liên kết