Bỏ qua tới nội dung
TakoAPI
Quay lại kỹ năng

markdown-browser

bởi 2233admin

Web & Frontend Development
317Lượt 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

Wrapper skill for OpenClaw web_fetch results.

Mô tả

Kỹ năng này làm gì

Post-processing wrapper for OpenClaw web_fetch results. Applies content policy decisions from Content-Signal headers, redacts sensitive path and query values from output URLs, and normalizes fetched content to markdown, HTML-fallback, or plain text. Outputs a stable schema for downstream agent logic without re-implementing network fetch.

Centralizes content policy, privacy redaction, and format normalization so individual agents don't each re-implement the same post-fetch logic.

Khi nào nên dùng

  • Checking AI-input policy before feeding a fetched page into an LLM
  • Stripping sensitive path and query values from URLs in agent logs
  • Converting HTML responses to markdown for text-based agent pipelines
  • Estimating token counts from fetched web content before processing
  • Standardizing web fetch output schema across multiple agent tools

Quy trình ví dụ

ĐẦU VÀO

User asks: fetch this documentation page and extract key points, respecting content policies

AGENT
  1. 1

    Call OpenClaw web_fetch to retrieve the target page

  2. 2

    Pass the result JSON and Content-Signal header to process_web_fetch_result

  3. 3

    Check policy_action: proceed if 'allow_input', halt if 'block_input'

  4. 4

    Use the normalized content field (markdown or text) for downstream summarization

  5. 5

    Log source_url with sensitive values already redacted

ĐẦU RA

Structured object with content, format, token_estimate, policy_action, and redacted source_url ready for agent consumption

Cài đặt

clawhub install 2233admin/markdown-browser

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

Liên kết