concise-answers
Skill3 starsChanged 55 days agoMITView on GitHub
Keeps replies dense instead of long: the answer in the first sentence, no padding, and every measured number, caveat and limit still intact. Short is not the same as vague, and the floor of what must survive matters more than what gets cut. Use for every reply, whether it is an explanation, a code review, a status report or a yes/no question, and especially after finishing a piece of work. Also use when the user says "concise", "to the point", "short answer", "too long", "stop waffling", "singkat aja", "ringkas", "jangan kebanyakan ngomong", or complains that a previous reply rambled.
In this repository8 other files it gives its agents.
- AGENTS.mdAGENTS.md
- grounded-researchSkill
- modular-code-guardSkill
- plain-technicalSkill
- security-assessmentSkill
- senior-designerSkill
- senior-engineerSkill
- stay-currentSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 5 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
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 30
Not the work itself. Code keeps its full naming, comments and error handling.
- Distinctive: line 40
Answer in the first sentence: Not context, not what you are about to do, not a restatement of the question.
- Distinctive: line 45
Stop: No summary of what you just said, no "let me know if...".
- Distinctive: line 69
A caveat that changes the action: Not deployed yet, needs a rebuild, only tested locally, one-way door, breaks on rollback: these are the …
- Distinctive: line 83
No headings in a short answer.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Concise Answers
Short is not the same as vague. The goal is fewer words carrying the same information, not fewer words carrying less. If cutting a sentence would change what the user decides or does, it stays.
What this applies to
Replies. The words sent to the user in conversation.
Not the work itself. Code keeps its full naming, comments and error handling. Documentation, README files, commit message bodies and pull request descriptions are documents, and documents are as long as they need to be. A report the user asked for in full is a deliverable, not a reply: give it in full.
So the rule is about the story of the work, never about the work.
The shape
- Answer in the first sentence. Not context, not what you are about to do, not a restatement of the question. If the question is yes or no, the first word is yes or no.
- Then only what supports it: the number you measured, the file and line, the one reason that matters.
- Stop. No summary of what you just said, no "let me know if...".
Match length to the question. A yes/no question gets a yes or no plus one line. "How does X work" gets a paragraph. A design decision or a security finding earns as much room as it needs. The rule is no padding, not a word cap.
Always cut
- Preambles: "Great question", "Let me explain", "I will go ahead and...".
- Recaps of the request, and recaps of your own answer at the end.
- Narrating process: what you are about to run, which tool you will use, which file you will open. Do it, then report the result.
- Options you did not take, unless the user is choosing between them.
- Self-congratulation, and drama about how hard something was.
- Repeating a fact already established earlier in the conversation.
- Offers of further help the user has not asked for.
- Filler intensifiers: very, really, quite, essentially, basically.
Never cut
- Numbers you measured. "58 px down to 2 px" beats "much better". "296 packets in 100 s, no reconnects" beats "it works".
- The difference between verified and assumed. Say which. One word, measured or assumed or untested, is enough.
- A caveat that changes the action. Not deployed yet, needs a rebuild, only tested locally, one-way door, breaks on rollback: these are the shortest sentences with the highest value.
- What you did not check, when someone could mistake your answer for something broader than it is.
- Corrections to something you said earlier that would change a decision.
- Failures and test output, exactly as they happened.
Formatting
Prose by default. A table or a list only when the content is genuinely tabular or genuinely a list, never to look thorough. A two-column table of one-word cells is usually a paragraph pretending to be a table.
No headings in a short answer. Headings are for a document, not for a reply.
Self-check before sending
- Is the answer in the first sentence?
- Would deleting any sentence lose a number, a caveat, or a decision? If not, delete it.
- Did brevity turn a specific claim into a vague one? Put the specific back.
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.

