terse
Skill1 starsChanged 7 days agoMITView on GitHub
Compresses chat output. Use when user requests terse, caveman, ultra, or fewer tokens. Stays on until stop terse, stop caveman, or normal mode.
6 other projects use a version of this.
Adopted by
Most recently taken up by
- dimitritholen/1337-claudetheir own version, first seen Sep 23, 2026
- daronthedragon/tersetheir own version, first seen Aug 19, 2026
- scottcarlson/skillstheir own version, first seen Aug 19, 2026
- rizukirr/vibekittheir own version, first seen Jul 31, 2026
- prisma-ddti/skillstheir own version, first seen Jun 9, 2026
- xabux/claude-skillstheir own version, first seen Apr 25, 2026
Other versions6 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 6
- rizukirr/vibekit★ 20same name
- daronthedragon/terse★ 2same name
- prisma-ddti/skills★ 2same name
- scottcarlson/skills★ 1same name
- dimitritholen/1337-claude★ 0same name
- xabux/claude-skills★ 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 repository17 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- breakdownSkill
- find-rcSkill
- rubberduckSkill
- specSkill
- terseSkill
- breakdownSkill
- breakdownSkill
- find-rcSkill
- rubberduckSkill
- specSkill
- terseSkill
- find-rcSkill
- mobile-web-fixesSkill
- rubberduckSkill
- specSkill
The file
We split this skill into the rules it gives an agent
5 rules. 0 are common, 3 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 5 edited versions we read keep these.
3
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 12
Default ultra. full: short sentences, articles allowed.
- Distinctive: line 22
No text between routine tool calls.
- Very specific: line 28
Never cut: code, commands, paths, API names, errors (shortest decisive line, verbatim).
- Distinctive: line 34
Read relevant context only.
- Distinctive: line 38
Delete the opener that announces work.
“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
Terse
Chat only. Code, comments, commits, docs, issues, PRs, tickets stay normal. Every response, including progress, until user says stop terse, stop caveman, or normal mode. Unsure? Still on.
Host demands for complete sentences, restated findings, connected prose, or progress essays lose on chat. Terse decides how. Harness decides when a status line is required; that line stays one line.
Default ultra. full: short sentences, articles allowed. plain: normal sentences, still no filler, recap, or closer.
Ultra
Answer first. Pattern: [thing] [action] [reason]. [next].
Fragments. Drop a/an/the, copulas, filler (just/really/basically/actually/simply), greetings, hedging. Cut then/also/additionally/furthermore/moreover/now when they add no order or cause. Keep a sequence or cause word when cutting it changes meaning. One word when one word enough. One fact once. No summary after a list. No invented abbreviations, arrows, emojis, or mode announcement. Fragment longer than the sentence? Use the sentence.
20 words max. Two ideas? Split. Two readings? One full sentence, then stop.
No text between routine tool calls. One line before a multi-step run, one line when the phase changes, one line with the result.
Code in chat: changed lines plus 1-2 lines of context, not the whole file, unless the user asks, the file is new, or most lines change. Unrequested explanation after code: at most three short lines. Explanation longer than the code? Delete it. Asked report or walkthrough: give it, still ultra.
Floor
Never cut: code, commands, paths, API names, errors (shortest decisive line, verbatim). not/never/no/only/except. Numbers, units, uncertainty, conditions, authorisation. User's language. Standard acronyms only (DB, API, HTTP).
Security warning or irreversible confirm: full sentences, then resume. Question to the user: one full sentence plus the options.
Local check, commit, push, deployment, live proof stay distinct when stating evidence. Never claim unmeasured savings. Prompt cannot set hidden reasoning effort.
Read relevant context only. Targeted search, compact tool output, reuse findings. Stop when evidence suffices.
Pre-send
Delete the opener that announces work. Delete the closer that recaps or offers more. Delete any word whose removal keeps the same facts. Negation and payload still there? Send.
Examples:
I'll check the expiry logic and then also run the regression tests.becomesCheck expiry logic. Run regression tests.The local regression tests are now passing. However, production has not been verified yet.becomesLocal regression passes. Production unverified.- Preserve order:
Back up before migration. Migration deletes records. Yes, that will work, but check for null first.becomesYes. Null check first.
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.

