sandbox-markdown
Parse markdown into a token stream inside a Code node or CodeAct action, with marked's lexer running in the guest
- License 许可
- AGPL-3.0
- 最近更新
- 2026-11-15
- Stars
- ★ 635
- Forks
- 91
01 前言 Front Matter
- name
- sandbox-markdown
- description
- Parse markdown into a token stream inside a Code node or CodeAct action, with marked's lexer running in the guest
02 资源文件 1
- 参考文档package.json875 B
03 SKILL.md 正文
Markdown in the sandbox
Specifier: @nodetool-ai/sandbox-markdown. One module, the marked root
export. Import it at the top of the body.
marked is pure JavaScript — no Node builtins, no DOM — so it compiles
straight into the QuickJS guest. There is no pre-built "extract headers"
helper; marked.lexer gives you the token stream and you read what you need
out of it.
lexer — markdown text to tokens
import { marked } from "@nodetool-ai/sandbox-markdown";
const tokens = marked.lexer(inputs.markdown);
return { tokenCount: tokens.length };
Each token carries a type (heading, paragraph, list, code, table,
link tokens nested inside paragraph/text tokens, etc.) and the fields
that type needs.
Headers → a document outline
import { marked } from "@nodetool-ai/sandbox-markdown";
const headers = marked.lexer(inputs.markdown)
.filter((t) => t.type === "heading")
.map((t) => ({ level: t.depth, text: t.text }));
return { headers };
Links → a flat list
marked nests inline tokens (links, emphasis) inside block tokens. Walk
tokens.links for a document's reference-style link definitions, or walk each
block's .tokens for inline links:
import { marked } from "@nodetool-ai/sandbox-markdown";
function collectLinks(tokens, out = []) {
for (const t of tokens) {
if (t.type === "link") out.push({ href: t.href, text: t.text });
if (t.tokens) collectLinks(t.tokens, out);
if (t.items) collectLinks(t.items, out);
}
return out;
}
const links = collectLinks(marked.lexer(inputs.markdown));
return { links };
Lists → nested items
import { marked } from "@nodetool-ai/sandbox-markdown";
const lists = marked.lexer(inputs.markdown)
.filter((t) => t.type === "list")
.map((t) => ({
ordered: t.ordered,
items: t.items.map((item) => item.text)
}));
return { lists };
Code blocks → language + source
import { marked } from "@nodetool-ai/sandbox-markdown";
const codeBlocks = marked.lexer(inputs.markdown)
.filter((t) => t.type === "code")
.map((t) => ({ language: t.lang || "text", code: t.text }));
return { codeBlocks };
Tables → rows of records
import { marked } from "@nodetool-ai/sandbox-markdown";
const tables = marked.lexer(inputs.markdown)
.filter((t) => t.type === "table")
.map((t) => {
const headers = t.header.map((cell) => cell.text);
return t.rows.map((row) =>
Object.fromEntries(row.map((cell, i) => [headers[i], cell.text]))
);
});
return { tables };
Markdown → HTML
import { marked } from "@nodetool-ai/sandbox-markdown";
const html = marked.parse(inputs.markdown);
return { html };
Gotchas
- Everything runs in the guest. The 64 MB guest heap holds your input text, the token tree, and the returned value at once.
marked.lexeris synchronous, unlikemarked.parse(which can be async when you register an async extension — the default configuration is not).- Token shapes differ by type. A
listtoken's items are themselves tokens with their own.tokens; atabletoken hasheader/rows, not a flatcellsarray. Log a sample token when a shape is unclear. - This pack is the only route to markdown parsing. There is no
data.parseMarkdownglobal; every library the sandbox offers is an importable module, and this is the one for markdown.
04 同仓库更多 Skills
humanizer
Rewrite AI-sounding text so it reads like the writer without changing what it says. Use when editing or reviewing prose for AI tells: not-X-but-Y contrasts, one-line closers, staged openers, forced triads, dashes everywhere, inflated claims, sales language, stock AI words, bold labels, or filler.
nodetool
Use the NodeTool MCP server to build, validate, run, and debug visual AI workflows, and to work with assets, collections, and jobs. Use when the user mentions NodeTool or asks for a workflow graph, a node, or a media generation run on their NodeTool server.
wan-2-6-prompting
Prompt Alibaba's Wan 2.6 video — text-to-video with a global style line plus timing-bracketed shots, and image-to-video that describes only what changes over time. Covers the shot-mode switch per provider, the prompt budget, negative prompts, motion intensity wording, audio input and prompt expansion. Use whenever the model id names a Wan 2.6 video route (alibaba/wan-2.6/text-to-video, alibaba/wan-2.6/image-to-video, wan-video/wan-2.6-t2v, wan-video/wan-2.6-i2v, wan-video/wan2.6-i2v-flash) on generate_video, animate_image or a TextToVideo node. Not for Wan 2.5 or 2.7, or the Wan 2.6 image routes.
video-workflow
Build a reusable NodeTool video workflow for batches or repeated inputs. For a single finished piece, use a storyboard skill.
video-clone
Analyze a reference video and recreate its shot structure with a new product or cast in NodeTool.
video-audio-continuity
Keep sound continuous across a multi-scene piece cut from generated video — why one clip per scene hard-cuts the audio at every boundary, when to write all the scenes into a single generation instead, the sound-design vocabulary that survives a provider's copyright filter, and how to check the length, the aspect ratio and the mix of what came back. Use when a piece has more than one scene and the video model writes its own audio (Seedance 2, Veo 3, MiniMax H3, Kling 2.6 and later), or when a cut's sound drops out at the shot changes.