notes
Skill1 starsChanged 33 days agoMITView on GitHub
Drop-in scratchpad. Renders the text you pass after `/notes` into a clean bulleted list under a timestamped "Your Notes" header. No editing, no reordering, no commentary — pure verbatim formatting so future-you can re-orient or discard at a glance, and know when you left it.
21 other projects use a version of this.
Adopted by
Most recently taken up by
- arch3rPro/ark-spacetheir own version, first seen Nov 8, 2026
- adeonir/agent-skillstheir own version, first seen Nov 6, 2026
- blogminhquy/javis-ostheir own version, first seen Nov 5, 2026
- letsloose501/sqs-agent-memorytheir own version, first seen Nov 3, 2026
- yourbeckett/beckett-for-claudetheir own version, first seen Nov 1, 2026
- baney75/myelinatetheir own version, first seen Oct 20, 2026
Other versions21 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 21
- blogminhquy/javis-os★ 711same name
- howells/fiction★ 23same name
- microsoft/PilotSwarm★ 14same name
- DaniilBrusnikUa/claude-code-handoff★ 9same name
- adeonir/agent-skills★ 8same name
- phanghonghao/THU-Awesome-Skills★ 8same name
- baney75/myelinate★ 6same name
- DoKu67/claude-code-skils★ 4same name
- jholhewres/devclaw-skills★ 4same name
- TTflysky/sirenhuisuo★ 4same name
11 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 repository56 other files it gives its agents.
- autopilotSkill
- before-afterSkill
- bigpictureSkill
- blast-radiusSkill
- brainstormSkill
- compoundSkill
- confidence-checkSkill
- contextualizeSkill
- create-ship-skillSkill
- create-skillSkill
- curateSkill
- debugSkill
- deep-researchSkill
- devtoolsSkill
- distillSkill
- done-checkSkill
- doneSkill
- double-checkSkill
- envsetSkill
- explainSkill
- fortifySkill
- goSkill
- handoffSkill
- harness-auditSkill
- interviewSkill
- matrixSkill
- mockupSkill
- mutateSkill
- postmortemSkill
- presentSkill
- qaSkill
- readableSkill
- recommendSkill
- reportSkill
- retireSkill
- review-planSkill
- scoutSkill
- sec-reviewSkill
- setup-ticketingSkill
- setup-worktreesSkill
- spec-checkSkill
- specSkill
- stepsSkill
- successionSkill
- super-reviewSkill
- tangentSkill
- tddSkill
- tend-memorySkill
- testSkill
- ticketSkill
- tldrSkill
- triageSkill
- waitSkill
- workSkill
- worth-itSkill
- wrap-upSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 2 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 15 edited versions we read keep these.
2
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 30
If there are no clear delimiters, render the whole thing as a single bullet.
- Very specific: line 32
If /notes was invoked with no arguments at all, output a single line: (no notes) — and stop.
- Very specific: line 36
No editing: Don't fix typos, don't expand abbreviations, don't rephrase for clarity.
- Very specific: line 38
No opinions: Do not add a recommendation, a "first I'd suggest…", or a synthesis bullet.
- Distinctive: line 40
No trailing commentary: Don't say "let me know if you want me to act on these." End at the last bullet.
“Common” is measured: kept by at least 6 in 10 of the 15 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
A self-contained notes pad. You invoke /notes <text> to leave yourself a reference drop — usually right before stepping away — so when you return, the first thing you see is your own pre-written context, untouched.
Output shape
A long literal divider line on its own line — at least 60 dashes (e.g. ------------------------------------------------------------) — followed by a blank line, then the header line, then the parsed bullets. Nothing else. No preamble, no signoff, no meta. The long dash string is intentional: a short markdown --- renders too short to read as a real separator.
The header line is Your Notes followed by an em-dash and the timestamp of this drop, e.g.:
**Your Notes** — Thu Jul 2, 2026 · 3:45 PMSince a note often sits parked for days, the stamp is the "when did I write this" anchor — it's why the timestamp lives right on the header where the eye lands first.
Timestamping
- Get the real wall-clock time with a single
datecall — do not guess it or reuse "today's date" from context (that has no time-of-day, and may be stale). Exact command:date '+%a %b %-d, %Y · %-I:%M %p'→ e.g.Thu Jul 2, 2026 · 3:45 PM. - This one
datecall is the only tool call the skill makes. If it somehow fails, fall back to the date from context with no time (Thu Jul 2, 2026) rather than inventing a time.
Parsing the notes
- Accept any reasonable delimiter you use. Likely shapes include: numbered (
1) ... 2) ...,1. ... 2. ...), hyphen/bullet (- ...,* ...), pipe-separated (note one | note two), slash-separated (note one / note two), double-dash (note one -- note two), semicolons, or just a single sentence with no delimiters. - Be generous — if you clearly intended multiple items separated by some character, split on it. Use judgment: a
/insidepath/to/fileis not a delimiter; a/surrounded by spaces between two clauses probably is. Same logic for-(avoid splitting hyphenated words or flags like--no-verify). - If there are no clear delimiters, render the whole thing as a single bullet.
- Render each parsed item as its own bullet, in the order you provided. Preserve the wording verbatim — no paraphrasing, no editorializing, no reordering, no added commentary, no helpful expansions. Preserve inline backticks, slash commands, filenames exactly as written.
- If
/noteswas invoked with no arguments at all, output a single line:(no notes)— and stop. Do not invent content, and skip thedatecall (nothing to stamp).
Hard rules
- No editing. Don't fix typos, don't expand abbreviations, don't rephrase for clarity. If you wrote
chk login thing, the bullet ischk login thing. - No reordering. Order is whatever you typed.
- No opinions. Do not add a recommendation, a "first I'd suggest…", or a synthesis bullet. This is not
/recommend— that exists separately if you want it. - No file writes, no investigation, and exactly one tool call — the
datestamp. Beyond that singledatecall, it's a pure read of the argument string → formatted output. No searching, no reading files, no acting on the notes' content. - No trailing commentary. Don't say "let me know if you want me to act on these." End at the last bullet.
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.

