tentacle-markdown2html

name	tentacle-markdown2html
description	Enterprise markdown-to-html renderer via API service. Sends markdown to api.tentacle.pro and writes returned html locally.

tentacle-markdown2html (Enterprise Edition)

Purpose

Call enterprise rendering service instead of local renderer:

• Endpoint: https://api.tentacle.pro/markdown2html
• Auth: API_KEY from .agents/skills/.env

Runtime

bun .agents/skills/tentacle-markdown2html/scripts/main.ts <markdown_file> [options]

Shared .env (.agents/skills/.env)

API_KEY=tpk-xxx

Input/Output

• Input: markdown file
• Request payload: { markdown, title? }
• Output: html file (default same-name .html)

Options

Flag	Description
--output <path>	Output HTML path
--title <text>	Title override
--summary <text>	Article summary (max 120 chars). Injected as summary: into frontmatter so the renderer displays it as a themed lead block after the h1. Use this when the markdown has no frontmatter or when the agent dynamically generates a summary.
--dry-run	Fetch only, skip file write

Preprocessing

• Before request, script normalizes Obsidian image embeds like ![[cover.jpg]] to standard markdown image syntax.
• This keeps enterprise behavior aligned with community flow and ensures downstream image extraction/replacement can work consistently.

WeChat-Oriented Expectation

• The generated HTML is expected to be consumed by tentacle-post2wechat, where:
  • inline images are uploaded with uploadimg semantics and rewritten as URL;
  • cover image is uploaded as permanent material and used as thumb_media_id.

Example

bun .agents/skills/tentacle-markdown2html/scripts/main.ts "02 Projects/article.md" \
  --summary "早餐是一个向量空间。煎饼、可丽饼和炒蛋安置在一个单纯形上，是否存在从未被探索的'暗早餐'？"