answer-first-structure
Skill0 starsChanged 4 months agoMITView on GitHub
Decide the bottom line before writing and put it in the first sentence, with only decision-relevant support after it. Use this whenever producing anything a reader will consume to decide or act — reports, emails, summaries, memos, recommendations, status updates, research findings, analysis writeups, PR descriptions, incident notes. Trigger especially on signals of a time-pressed reader ("my VP has 30 seconds", "keep it brief", "TL;DR", "what should we do?"), but also when no such signal is given — readers are always more time-pressed than the writer assumes, so apply it to any question that has an answer or any task that ends in a recommendation.
In this repository10 other files it gives its agents.
- assess-before-actingSkill
- assumption-auditSkill
- calibrated-claimsSkill
- finish-the-turnSkill
- grounded-reportingSkill
- plan-before-executingSkill
- root-cause-before-fixingSkill
- rubric-self-checkSkill
- scope-disciplineSkill
- verify-before-answeringSkill
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 12
Before drafting anything, answer the reader's question in one sentence.
- Distinctive: line 13
Commit to that sentence and make it the opening line of the output.
- Distinctive: line 26
Never narrate your own diligence ("I carefully verified...", "having thoroughly checked...", "as instructed, I...").
- Distinctive: line 28
Keep length proportionate to the task: a one-question email gets a few sentences.
- Distinctive: line 29
If the honest bottom line is uncertain, say that in the first sentence too ("Inconclusive — here's what would settle it") rather than …
≈ “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
Answer-First Structure
Readers decide in the first seconds whether to keep reading, and many stop there. If the bottom line arrives last, the time-pressed reader acts on an incomplete picture — or doesn't act at all. Writing the answer first also disciplines the writer: you cannot state a conclusion in one sentence until you have actually reached one, so weak or hedged thinking surfaces before the reader sees it, not after.
Process
- Before drafting anything, answer the reader's question in one sentence. If you can't, the analysis isn't done — finish it first. This is the checkpoint: no conclusion, no draft.
- Commit to that sentence and make it the opening line of the output. Not context, not background, not what you did — the answer or recommendation itself.
- Build the structure under it: pick the 2–4 points the reader needs to trust or act on the conclusion, ordered by decision relevance, not by the order you discovered them.
- Cut everything that supports your process rather than the reader's decision — how you searched, what you ruled out along the way, how careful you were. If a caveat changes the decision, keep it and put it early; otherwise drop it.
- Reread only the first sentence and ask: does a reader who stops here take the right action? If not, rewrite the first sentence, not the rest.
What this looks like
- Email to a busy exec: "Recommend we delay the launch two weeks; the payment vendor won't be certified before the 20th." Then two lines of support, not a timeline of meetings.
- Research summary: "The evidence favors option B: it's cheaper at our scale and the main risk (vendor lock-in) has a workable exit." Sources and comparison follow for those who read on.
- PR description or incident note: "Fixes the checkout timeout by raising the connection-pool cap; root cause was pool exhaustion under concurrent sessions." Details of the investigation only as far as a reviewer needs them.
Output rules
- Never narrate your own diligence ("I carefully verified...", "having thoroughly checked...", "as instructed, I..."). The discipline shows in the work product; describing it spends the reader's time proving nothing.
- The checkpoint work — reaching the conclusion, weighing caveats — happens before or while producing the answer. Only its results appear in the output, and only where they change what the reader does.
- Keep length proportionate to the task: a one-question email gets a few sentences; a decision memo gets a page. Support may be as deep as the decision warrants, but it always comes after the answer.
- If the honest bottom line is uncertain, say that in the first sentence too ("Inconclusive — here's what would settle it") rather than burying the uncertainty under structure that implies confidence.
Grounding: BLUF (US Army AR 25-50); Barbara Minto, The Pyramid Principle (answer-first structure); Anthropic, Prompting Claude Fable 5 ("Lead with the outcome").
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.

