huh
Skill2 starsChanged 43 days agoMITView on GitHub
Now, next, you. Wait, where were we? Say the last message again as three labelled lines - where it stands, what happens without me, and what you need from me. For picking a thread back up.
7 other projects use a version of this.
Adopted by
Most recently taken up by
- andresthor/skillstheir own version, first seen Nov 1, 2026
- MichaelHolley/cc-pluginstheir own version, first seen Oct 25, 2026
- adam-paterson/skillstheir own version, first seen Oct 11, 2026
- nielsmadan/skillstheir own version, first seen Sep 5, 2026
- timharris707/skillstheir own version, first seen Sep 5, 2026
- blixvip/skillstheir own version, first seen Aug 19, 2026
Other versions7 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 7
- timharris707/skills★ 44same name
- blixvip/skills★ 4same name
- andresthor/skills★ 2same name
- joshuatownsend/huh-skill★ 2same name
- adam-paterson/skills★ 1same name
- nielsmadan/skills★ 1same name
- MichaelHolley/cc-plugins★ 0same name
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 repository16 other files it gives its agents.
- audit-backendSkill
- audit-frontendSkill
- audit-securitySkill
- ralph-loopSkill
- ralph-planSkill
- againSkill
- elimSkill
- ughSkill
- code-audit-jevSkill
- jev-email-analysisSkill
- prompt-injection-evalSkill
- resume-mirror-evalSkill
- run-jevSkill
- traffic-triage-evalSkill
- built-withSkill
- github-remediate-vulnsSkill
The file
We split this skill into the rules it gives an agent
6 rules. 0 are common, 4 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
4
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 25
Write each label in bold - Now:, Next:, You: - and the urgency prefix on each bullet too.
- Distinctive: line 32
No file paths, tool names, commands or shop talk.
- Very specific: line 68
When every ask shares one urgency, say it once on the You: line and drop the per-bullet prefixes.
- Distinctive: line 80
Keep the ask, and never invent one.
- Distinctive: line 85
No headings, tables or code beyond the three labels, and no bullets except the multiple-ask case above.
- Distinctive: line 88
If the message was already this short, say so and leave it alone.
“Common” is measured: kept by at least 6 in 10 of the 5 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
Say your last message again for someone who is tired. They have nine things open and are picking this one back up.
Shape
Three labelled lines, nothing before or after them:
Now: where it stands.
Next: what happens without them.
You: the ask, or "nothing".One or two sentences a line. Keep a label even when its line is empty - a missing label reads as forgotten, not as clear.
Write each label in bold - **Now:**, **Next:**, **You:** - and the urgency
prefix on each bullet too. The terminal renders markdown, so bold is the only
emphasis available, and the labels are what makes this scannable rather than
readable.
Words
No file paths, tool names, commands or shop talk. Say what a thing does, not what it is called. Short common words: use, not utilize. Wrong, not incorrect. Left out, not omitted.
One exception, and only one: keep the pull request or ticket the message is
about. Once, as a short link - [#1733](the url) - on the thing it names. It
does not count against the shape.
The ask
An ask is something only they can do: a decision, an approval, an action, an answer. "I will ping you when it lands" is not an ask.
"Nothing" and an ask cannot share a line. If there is an ask, it comes first. Say how soon it is needed after it, never before.
One ask stays inline on the You: line. Two or more become bullets under it,
one per ask, so the count is visible before a word is read. Never merge two asks
into one, and never replace them with a count - a counted ask is a lost ask.
Order the bullets by what is needed soonest.
Each bullet leads with its urgency, so the reader can find tonight's work without reading the topics:
You:
- Before I merge: both changes now, or hold the app side until after the campaign?
- No rush: file a check that catches the next one?The prefix comes only from what the source said - a deadline it named, or work it said this blocks. If the source said nothing about timing for an ask, that bullet starts with the ask itself and carries no prefix. Never invent urgency, and never invent a deadline.
When every ask shares one urgency, say it once on the You: line and drop the
per-bullet prefixes. A prefix repeated on every bullet is not information, and
"Also before starting" is a clumsy way of saying "same as the one above":
You: both before any work starts.
- Cross-agent for all nine, which means ...
- Does your local copy stay the master ...Faithfulness
Keep the ask, and never invent one. Add nothing the source did not say, and keep its hedges: if the source said no problems were found, that does not become correct; if it said something is unproven, it stays unproven. Keep every number the source gave.
No headings, tables or code beyond the three labels, and no bullets except the multiple-ask case above.
If the message was already this short, say so and leave it alone.
Sibling: /ugh (one breath, the tldr) shares everything above except the Shape section. A
rule changed here belongs there too.
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.

