Bỏ qua tới nội dung
TakoAPI
Quay lại kỹ năng

macarena-test

bởi misirov

Coding Agents & IDEs
1.1kLượt 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 threat model for OpenClaw gateway hosts.

Mô tả

Kỹ năng này làm gì

Clawstrike audits an OpenClaw gateway deployment for misconfigurations and real-world attack paths. It runs a verified collection script, evaluates configuration, network exposure, filesystem hygiene, and installed skills/plugins, then produces an OK/VULNERABLE report with severity ratings, evidence excerpts, and fix instructions.

Produces a structured, evidence-backed report with automatic secret redaction instead of requiring manual ad-hoc checks across config files, network state, and plugin directories.

Khi nào nên dùng

  • Checking whether a self-hosted OpenClaw gateway is reachable without authentication
  • Auditing installed skills and plugins for supply chain risks before a team deployment
  • Verifying firewall and port exposure on an OpenClaw host
  • Reviewing filesystem permissions and SUID/SGID binaries on the gateway machine
  • Generating a shareable security report before granting colleagues access to an OpenClaw instance

Quy trình ví dụ

ĐẦU VÀO

User asks: audit my OpenClaw deployment and tell me if anything is vulnerable

AGENT
  1. 1

    Run scripts/collect_verified.sh in the current working directory to gather system evidence

  2. 2

    Read verified-bundle.json to obtain timestamp, OS, OpenClaw version, config path, and runtime context

  3. 3

    Evaluate every required check from references/required-checks.md against evidence in the bundle

  4. 4

    Build a concise threat model aligned with the findings using references/threat-model.md

  5. 5

    Emit a findings table with severity, evidence excerpts, and fix instructions for each issue

ĐẦU RA

A structured OK/VULNERABLE report listing each check result, severity level, redacted evidence, and remediation steps

Cài đặt

clawhub install misirov/macarena-test

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

Liên kết