Bỏ qua tới nội dung
TakoAPI
Quay lại kỹ năng

clawgatesecure

bởi thestormshadow

Productivity & Tasks
—Chưa có đánh giá
Đánh giá của bạn

Tóm tắt

Advanced security protocol for LLM agents focusing.

Mô tả

Kỹ năng này làm gì

Security protocol for LLM agents that enforces prompt injection mitigation, mandatory skill auditing, and data exfiltration prevention. All external input is treated as potentially malicious and passed through a scrubber and sandbox before execution. Agent output is scanned for secrets, PII, and configured keys before delivery.

Unlike ad-hoc validation, it enforces an immutable execution lock that halts the agent until a human explicitly approves the audit report, preventing automatic action on untrusted input.

Khi nào nên dùng

  • Auditing a third-party skill file before activating it in the agent
  • Scanning agent output for accidentally exposed API keys or tokens
  • Blocking prompt injection attempts from web-scraped or user-provided content
  • Encrypting sensitive agent memories stored at rest
  • Generating a risk score report before approving a multi-step workflow

Quy trình ví dụ

ĐẦU VÀO

User asks: Activate this new skill I downloaded from an external source

AGENT
  1. 1

    Scrubber sanitizes the skill file, stripping scripts and hidden metadata

  2. 2

    Sandbox sub-agent analyzes the file with zero tools and zero memory access

  3. 3

    Agent performs line-by-line audit checking for exfiltration vectors, mining code, and backdoors

  4. 4

    Unified audit report is generated with a Global Risk Score (1-10) and dependency map

  5. 5

    Execution locks until user sends an unequivocal affirmative command to proceed

ĐẦU RA

Audit report delivered with risk score and findings; skill activates only after explicit user approval

Cài đặt

clawhub install thestormshadow/clawgatesecure

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

Liên kết