Trafilatura Web Extract

Incubating Butler wrapper spec for upstream 'Trafilatura'.

ocean326 f3d9232 3 files · 3.5 KB Updated

File contents

Trafilatura

Status

  • incubating
  • 这不是生产可执行 skill,而是 Butler 对外部上游的包装草案。
  • 默认不得加入 chat_default / codex_default / automation_safe

Proposed Butler Shape

  • web article extraction, cleanup, metadata capture, source note generation

Upstream

Why This Matters

  • Good fit for turning arbitrary pages into clean text for downstream summarization and archival.
  • Works well as a shared primitive under web-note, forum snapshot, and research ingestion skills.

Risks

  • Dynamic sites and login walls still need browser automation or a secondary fetch path.
  • Needs source attribution rules in Butler output.

Suggested Future Exposure

  • chat_content_share
  • codex_default

Implementation Checklist

  • 明确 Butler 输入输出 contract,不直接复刻上游 API。
  • 确认认证、限流、缓存和错误恢复策略。
  • 补执行脚本或 tool bridge,而不是只停留在说明文档。
  • 补测试与 skill-pool-verify / skill-pool-maintain 校验。
  • 审阅后再决定是否加入 collection registry。

ocean326/butler/tree/main/butler_main/platform/skills/pool/incubating/general/trafilatura-web-extract commit f3d9232f04

Frequently asked questions

npx skillmds@latest add ocean326/trafilatura-web-extract