notion
Skill9 starsChanged 5 months agoMITView on GitHub
Ingest a Notion note into notes/ — live via the Notion MCP (page URL, page ID, or title to search) or from a local .md/.html export
100 other projects use a version of this.
Adopted by
Most recently taken up by
- TucsenRoman/claude-skillstheir own version, first seen Nov 16, 2026
- zkManuel0123/CodexClawtheir own version, first seen Nov 16, 2026
- metaspartan/cybaratheir own version, first seen Nov 15, 2026
- AceDataCloud/Skillstheir own version, first seen Nov 8, 2026
- ashleysbrain/skillstheir own version, first seen Nov 8, 2026
- sidmanale643/pekka-bottheir own version, first seen Nov 8, 2026
Used with
Projects with it also use taskflow-inbox-triage, food-order and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
AtomicBot-aiSkilltaskflow-inbox-triage
Also in 6 projects that use this
seedprodSkillfood-order
Reorder Foodora orders + track ETA/status with ordercli. Never confirm without explicit user approval. Triggers: order food, reorder, track ETA.
Also in 6 projects that use this
seedprodSkillvoice-call
Start voice calls via the OpenClaw voice-call plugin.
Also in 19 projects that use this
jiulingyunSkillBlueBubbles
为Clawdbot构建或更新BlueBubbles外部通道插件(扩展包,REST发送/探测,webhook入站)。
Also in 19 projects that use this
hkyutongSkillopenclaw-release-maintainer
Maintainer workflow for OpenClaw releases, prereleases, changelog release notes, and publish validation. Use when Codex needs to prepare or verify stable or beta release steps, align version naming, assemble release notes, check release auth requirements, or validate publish-time commands and artifacts.
Also in 4 projects that use this
AtomicBot-aiSkilltaskflow
Also in 6 projects that use this
- +6 more from seedprod
Other versions100 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 100
- molt-bot/clawdbot★ 395ksame name
- steipete/warelay★ 395ksame name
- moltbot/clawdbot★ 395ksame name
- openclaw/openclaw★ 395ksame name
- steipete/clawdis★ 394ksame name
- moltbot/moltbot★ 394ksame name
- clawdbot/clawdbot★ 394ksame name
- spinabot/brigade★ 11ksame name
- the-open-agent/openagent★ 5.7ksame name
- casibase/casibase★ 5.7ksame name
90 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository14 other files it gives its agents.
- .claude/AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- alphaxivSkill
- askSkill
- autoexploreSkill
- collateSkill
- curriculumSkill
- editSkill
- goalSkill
- obsidianSkill
- retrieveSkill
- tutorSkill
- webSkill
- zoteroSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
13 rules. 0 are common, 7 are distinctive, and 6 are very specific to this project.
0
Common
Most of the 95 edited versions we read keep these.
7
Distinctive
This author's own choices. Worth a look.
6
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 21
If $ARGUMENTS is empty, ask the user for a Notion URL, page ID, title, or local export path, then continue.
- Very specific: line 47
Search the workspace and confirm the page with the user before ingesting (same pattern as /alphaxiv's title search):
- Very specific: line 56
Ask the user which one to ingest (use AskUserQuestion).
- Distinctive: line 57
If search returns nothing, tell the user and ask for a URL/ID or a different title.
- Distinctive: line 75
Read properties from the page's property block (Tags, Status, Category, multi-select, etc.), not from the prose.
- Very specific: line 77
Use the readable page body as the content.
- Distinctive: line 87
Read the export at $ARGUMENTS and pull out title, properties, and body.
- Very specific: line 170
Use [] only if the note has no discernible topics.
- Distinctive: line 173
Before writing, verify the envelope:
- Distinctive: line 184
If any check fails, fix the envelope before writing.
- Distinctive: line 194
After writing, suggest the user:
- Very specific: line 196
Run /retrieve <topic> to find related sources and existing pages.
- Very specific: line 197
Run /collate <topic> to synthesise a paper or concept page from this note together with other sources on the same topic.
“Common” is measured: kept by at least 6 in 10 of the 95 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Ingest a Notion note — a personal note — and produce a structured
notes/notion_<slug>.md file following the project template. This is the Notion
sibling of /obsidian; it is normally invoked by /collate as its ingestion
step, but can be run directly.
There are two input modes, both converging on the same extract → format → write:
- Live (Notion MCP) — preferred. $ARGUMENTS is a Notion page URL, a page
ID, or a title to search for. The page is pulled from the live workspace with
mcp__notion__notion-fetch(andmcp__notion__notion-searchto resolve a title). - Local export — $ARGUMENTS is a path to a
.mdor.htmlNotion export (wherever you keep it on your filesystem). Parsed directly, no MCP. This is the original no-MCP path and the fallback when the workspace isn't reachable.
If $ARGUMENTS is empty, ask the user for a Notion URL, page ID, title, or local export path, then continue.
Layout
Output goes to notes/ at the repo root (personal notes, as opposed to objective
sources under sources/). The template is .claude/Templates/note.md and is authoritative
for the schema; you may read an existing notes/notion_*.md for style, but follow
the schema below over any older file.
Step 0 — Route the input
Classify $ARGUMENTS and pick the mode. Check in this order:
- Notion URL — contains
notion.so/or.notion.site/→ MCP fetch (Step 1M), passing the URL straight tonotion-fetch. - Local export — ends in
.mdor.html(case-insensitive), or is an existing path on disk (confirm withGlob/Read) → Local export (Step 1L). - Page ID — a bare 32-hex string, with or without UUID dashes (e.g.
1a2b3c4d5e6f7890abcdef0123456789or1a2b3c4d-5e6f-7890-abcd-ef0123456789) → MCP fetch (Step 1M), passing the ID tonotion-fetch. - Title / topic — anything else → MCP search (Step 1S): search, confirm, then fetch.
Step 1S — Resolve a title via search (MCP search mode)
Search the workspace and confirm the page with the user before ingesting (same
pattern as /alphaxiv's title search):
mcp__notion__notion-search query="$ARGUMENTS" query_type="internal" page_size=5- Show the top matches as a short numbered list — title + URL (and the parent / last-edited hint if present) for each.
- Ask the user which one to ingest (use
AskUserQuestion). Do not auto-pick. - If search returns nothing, tell the user and ask for a URL/ID or a different title.
Once the user confirms a match, take its page URL (or ID) and continue to Step 1M.
Step 1M — Fetch the page (MCP fetch mode)
Fetch the page content from the live workspace:
mcp__notion__notion-fetch id="<the URL or page ID>"notion-fetch returns the page in Notion-flavored Markdown wrapped in an envelope
(e.g. a <page url="..."> tag with the page's properties and body; child databases
appear as <data-source> markers). When extracting in Step 2:
- Take the canonical page URL from the fetch envelope — this becomes the single
source_linksentry. - Read properties from the page's property block (Tags, Status, Category, multi-select, etc.), not from the prose.
- Use the readable page body as the content; do not copy the envelope tags,
the raw property block, or internal Notion block IDs into
## Content. - Preserve Notion structures as Markdown: headings, lists, to-do checkboxes,
callouts, toggles, quotes, code blocks, tables, equations, and any internal links
(relative
.mdlinks orhttps://www.notion.so/...URLs) verbatim.
Then go to Step 2.
Step 1L — Read the export (local export mode)
Read the export at $ARGUMENTS and pull out title, properties, and body. Route by extension:
.md— useReaddirectly; Notion markdown is clean..html— strip tags to text first:python3 -c " import sys from pathlib import Path from html.parser import HTMLParser class T(HTMLParser): def __init__(self): super().__init__(); self.out=[] def handle_data(self,d): self.out.append(d) p=T(); p.feed(Path(sys.argv[1]).read_text(encoding='utf-8')) sys.stdout.write(''.join(p.out)) " "$ARGUMENTS"
Local Notion exports carry a few Notion-isms the live MCP path doesn't:
- Filenames and the H1 carry a trailing 32-hex page ID in real exports, e.g.
Transformers Notes 1a2b3c4d5e6f7890abcdef0123456789.mdand a matching# Transformers Notes 1a2b3c4d.... Strip that trailing hex blob from both. - A leading property block often follows the H1 — either
Key: valuelines or a small markdown table — holding Notion page properties (Tags, Status, Category, multi-select). It is metadata, not prose. - Internal links appear as relative
.mdlinks ([Name](Some%20Page%20<hex>.md)) orhttps://www.notion.so/...URLs. Preserve these verbatim. - The fixtures in this repo are clean notes with none of the above — handle that case too (no markers ⇒ derive concepts from the content alone).
For local mode the source_links entry is $ARGUMENTS verbatim. Then go to Step 2.
Step 2 — Extract
From whichever mode produced the page, identify:
title—- MCP: the page title from the fetch envelope. Strip any trailing 32-hex ID if one leaked into it.
- Local: the H1 (
# ...) with any trailing 32-hex page ID stripped; if there is no H1, fall back to the filename stem (strip the trailing hex ID, replace_/-/%20with spaces) in Title Case.
- Properties — the page's property block (Tags, Status, Category, multi-select,
keywords). Collect the values of any tag-like property, splitting
comma-separated values. These are metadata — do not copy the raw property block
into
## Content. - Wikilinks and tags — a Notion note may contain
[[wikilinks]]and#tags(someone who dual-uses Obsidian, or types them by hand). When present, treat them exactly as/obsidiandoes: collect the target of each[[wikilink]](the part before|if aliased, trimmed) and each#tagname (leading#stripped), and leave them in place inside the body — the navigator uses[[wikilinks]]as graph edges, so they must be preserved verbatim. - Body — everything after the title and property block. Preserve the author's
wording, headings, lists, code blocks, emphasis,
[[wikilinks]],#tags, and any internal links/URLs verbatim. Do not summarise unless the note is very long. - Language — if body prose is in Vietnamese (or mixed), translate it to concise
English preserving the original meaning; prose already in English is left unchanged,
and code, math, identifiers, URLs,
[[wikilinks]], and#tagsstay verbatim. - Length — if the note is too long, summarise it to keep it concise while keeping
the original meaning; preserve structure, math, links,
[[wikilinks]], and#tags.
Step 3 — Format
Fill in the notion schema below. It extends the shared .claude/Templates/note.md (type,
title, concepts_mentioned, source_links); the schema here is authoritative for
notion notes.
type: notion_notetitle: from Step 2.source_links: a YAML block list with a single entry:- MCP: the canonical Notion page URL from Step 1M.
Local: $ARGUMENTS verbatim.
source_links: - <Notion page URL, or $ARGUMENTS for a local export>
concepts_mentioned: an inline YAML list built from three sources, in this order, deduped at exact-string level preserving first occurrence:- the
[[wikilink]]targets and#tagnames from Step 2 (if any); - the tag-like property values from Step 2 (if any);
- the key topics named in the body — identified by judgment, since clean Notion
notes carry no markers (e.g.
[transformers, self-attention, RLHF]). Use[]only if the note has no discernible topics.
- the
## Content: the body from Step 2, faithfully preserved.
Before writing, verify the envelope:
- Starts with
---on line 1; has a closing---delimiter. - Contains exactly these four frontmatter keys:
type,title,source_links,concepts_mentioned. typeis the literal stringnotion_note.source_linksis a YAML list whose single entry is the Notion page URL (MCP mode) or $ARGUMENTS verbatim (local mode).concepts_mentionedis a YAML list (possibly empty).- Body contains a non-empty
## Contentsection.
If any check fails, fix the envelope before writing.
Step 4 — Write
Slugify the title to snake_case (lowercase, spaces/dashes → underscores, drop
non-alphanumeric characters). Write to notes/notion_<slug>.md with the Write
tool. If the file already exists, ask the user before overwriting.
Step 5 — Propose
After writing, suggest the user:
- Run
/retrieve <topic>to find related sources and existing pages. - Run
/collate <topic>to synthesise a paper or concept page from this note together with other sources on the same topic.
Wrap up
Report: output file path, the title, the input mode used (live MCP or local export)
and the resolved page URL/path, the concepts extracted (noting whether they came from
[[wikilinks]]/#tags, properties, or the body), any [[wikilinks]]/#tags
preserved, and anything skipped (one-line reason each).
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

