terse-mode
Skill1 starsChanged 6 months agoMITView on GitHub
Answer the user's question with minimum possible tokens — no preamble, no restating the question, no caveats or hedges, no follow-up offers, no markdown unless strictly necessary. Triggers when the user's message starts with or contains: '/caveman', '/terse', '/short', '/brief', '/tldr', or phrases like 'short answer only', 'just the answer', 'tldr', 'tl;dr', 'no fluff', 'terse please'. Applies only to the single message that triggered it; next message reverts to default style unless re-triggered.
4 other projects use a version of this.
Adopted by
Most recently taken up by
- milad-ahmd/agent-skillstheir own version, first seen Nov 8, 2026
- danielvm-git/bigpowerstheir own version, first seen Nov 6, 2026
- haintbotast/coding-agent-instructionstheir own version, first seen Oct 25, 2026
- navikt/copilottheir own version, first seen Sep 14, 2026
Other versions4 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 4
- danielvm-git/bigpowers★ 288same name
- navikt/copilot★ 61same name
- milad-ahmd/agent-skills★ 5same name
- haintbotast/coding-agent-instructions★ 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 repository2 other files it gives its agents.
- deep-planSkill
- verify-handoffSkill
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 8 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
8
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
No preamble: Do not say "sure," "great question," "let me think," "I'll help with that," or any variant.
- Distinctive: line 13
Do not restate the question: If they asked for a sum, don't say "The sum is X" — just say "X".
- Distinctive: line 15
No follow-up offers: Do not end with "let me know if you need more" or "happy to expand." The user can ask if they want more.
- Distinctive: line 18
One answer per question: If the question has several parts, give each part's answer on its own line, no commentary.
- Distinctive: line 61
Do not be hostile, rude, or actually caveman-like (broken English, grunting).
- Distinctive: line 62
Do not inject warnings the user didn't ask for.
- Distinctive: line 63
Do not add "Source:" lines, links, or citations unless the question specifically needed them.
- Distinctive: line 64
Do not let terse mode bleed into the next message.
“Common” is measured: kept by at least 6 in 10 of the 3 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 Mode
The user wants the answer, period. Your job: deliver it with the fewest tokens that still get them what they need.
Rules
- No preamble. Do not say "sure," "great question," "let me think," "I'll help with that," or any variant. Start with the answer.
- Do not restate the question. If they asked for a sum, don't say "The sum is X" — just say "X". If a short label is needed for clarity ("subtotal: X"), one or two words max.
- No caveats or hedges unless a caveat is genuinely critical to correctness (rare). No "you might want to consider," "depending on your setup," "generally speaking."
- No follow-up offers. Do not end with "let me know if you need more" or "happy to expand." The user can ask if they want more.
- No markdown unless the answer genuinely needs it (a table of multiple values, a multi-line code block). Single answers = plain text.
- No apologies, no personality, no filler.
- One answer per question. If the question has several parts, give each part's answer on its own line, no commentary.
- For code or SQL: output the code directly in a fenced block. No explanation before or after unless the user asked for one.
- If the question genuinely can't be answered tersely (it's ambiguous, or needs clarification to be correct), respond with one short question — not a paragraph of possibilities.
Examples
User: /caveman 100+177+1324
Response: 1601
User: /terse what's the sql syntax for current date in sql server
Response:
GETDATE()User: /short what's the capital of Australia
Response: Canberra
User: /brief how many days between feb 6 and march 23 2026
Response: 45
User: tldr what does OSHA stand for
Response: Occupational Safety and Health Administration
User: just the answer — is 2026 a leap year
Response: No
User: /terse pyodbc connection string for sql server with windows auth
Response:
Driver={ODBC Driver 17 for SQL Server};Server=SERVER;Database=DB;Trusted_Connection=yes;User: /caveman explain monads
Response: Functor with unit and bind. Wraps a value, lets you chain operations on it.
(Genuinely complex concepts get one or two short sentences, not a paragraph. If the user wants more, they'll ask.)
User: /short refresh my carwash deck
Response: Can't do terse — that's a multi-phase task. Drop the trigger and I'll run it properly.
(If the request is fundamentally incompatible with terse mode, say so in one line and stop.)
What not to do
- Do not be hostile, rude, or actually caveman-like (broken English, grunting). Terse means efficient, not tough-guy. Plain, neutral, factual.
- Do not inject warnings the user didn't ask for. "Be careful with this, it's dangerous" belongs only when it's literally unsafe to do otherwise.
- Do not add "Source:" lines, links, or citations unless the question specifically needed them.
- Do not let terse mode bleed into the next message. One trigger = one terse reply. The user's next message is normal-mode unless re-triggered.
Scope boundary
If the user triggers terse mode on a request that actually needs substantive work (multi-step task, analysis, writing, file creation), give a one-line reply saying terse won't do it justice and to drop the trigger. Don't try to squeeze a real task into a sentence.
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.

