agentleFS
mojopolymojopoly/three-loops/skills/interrogate/SKILL.md

interrogate

Skill0 starsChanged 3 months agoMITView on GitHub

Interview the user about a plan, feature, or design until every branch that would change the build is resolved. Use before generating any non-trivial feature, when requirements are vague, or when the user asks to stress-test a plan.

73 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use ketchup-video, principle-explain-the-number and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • tristanphvnSkill

    ketchup-video

    Catch the user up like /keel:ketchup, with a short silent video of slides for what happened and the waiting calls as text under it. Use for /keel:ketchup-video.

    Also in 14 projects that use this

  • ianclearySkill

    principle-explain-the-number

    Apply before you trust, report, or act on a number you measured: a speedup, a regression, a throughput, a latency, or an eval result. Find what limits it, and rule out that it measured something other than the work you think.

    Also in 29 projects that use this

  • eryzerzSkill

    make-bot-ui

    Use when building a custom UI (page, dashboard, buttons) that should wake a Grok Bot over a webhook, when the user must provide a webhook sender key, or when exposing that UI on Tailscale.

    Also in 20 projects that use this

  • bnemaSkill

    setup-pstack

    Configure which Pi models pstack uses per role. Detects available provider/model specs and writes pstack's user configuration. Use for /setup-pstack, "configure pstack models", or changing pstack model choices.

    Also in 25 projects that use this

  • ianclearySkill

    poteto-help

    Guides users through pstack setup, /poteto-mode, and picking the skill, playbook, or principle for a task. Type /poteto-help with a question.

    Also in 14 projects that use this

  • Xia-AtaraxiaSkill

    benchmark-checklist

    Vets performance measurements for the limiter, tuning, physical limits, errors, repeatability, relevance, and proof that timed work happened. It applies to 'check this benchmark', 'report this speedup', measured regressions, and choosing a library or configuration from numbers. It reports spread and gaps before drawing a conclusion. Not for general testing policy or choosing test oracles - use principle-testing.

    Also in 37 projects that use this

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 5 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 73 edited versions we read keep these.

  • 5

    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
  1. Distinctive:

    Ask in small batches (1–3 questions), hardest first.

    line 14
  2. Distinctive:

    Prefer concrete failure scenarios over abstract preferences: "the worker dies after the row is claimed but before the job is queued — what …

    line 15
  3. Distinctive:

    After each answer, update the open-branch list and say what it closed.

    line 18
  4. Distinctive:

    Start from whatever material exists (notes, threads, a spike) and ask only what it doesn't answer.

    line 22
  5. Distinctive:

    Stop when the remaining open branches would not change what gets built — list them as "deferred" with an owner.

    line 47

“Common” is measured: kept by at least 6 in 10 of the 73 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Interrogate

Generation amplifies ambiguity: a vague request produces a confident, plausible, wrong implementation. This skill resolves the ambiguity before any code exists, where a question costs seconds instead of a review round.

Rules of engagement

  • Ask in small batches (1–3 questions), hardest first. Never a wall of 15.
  • Prefer concrete failure scenarios over abstract preferences: "the worker dies after the row is claimed but before the job is queued — what should the user see?" beats "how should we handle errors?"
  • After each answer, update the open-branch list and say what it closed.
  • Offer a recommendation with each question when you have one — for open decisions, as numbered options with the recommendation marked and its reason stated. The user should mostly be choosing, not authoring.
  • Start from whatever material exists (notes, threads, a spike) and ask only what it doesn't answer; arriving with nothing is equally fine.
  • Push back once when an answer creates a contradiction with an earlier one or with the codebase. Twice is nagging; note it and move on.

Question categories — walk all that apply

  1. Scope edges — what is explicitly OUT? What is the smallest version that would still be worth shipping?
  2. State and concurrency — what are the states, who moves them, and what happens when two movers race? (Cancel-vs-fire, retry-vs-supersede, double-click, two tabs.)
  3. Time — timezones, DST, clock skew, "what does 'tomorrow' mean", staleness cutoffs, how late is too late.
  4. Failure policy — for each external call: retry, fail loudly, or degrade? Who is told, in what words?
  5. Data lifecycle — creation, mutation ownership, deletion. What happens to dependent rows when the parent goes away?
  6. Permissions — who can trigger this? What about admins acting as users?
  7. Observability — how will you know it worked in production? What log line or metric proves it?
  8. Blast radius — what existing behavior could this change accidentally?

Stop condition

Stop when the remaining open branches would not change what gets built — list them as "deferred" with an owner. Do not stop because the user got tired; compress instead: batch the remaining branches with your recommended answer for each and ask for line-item vetoes.

Exit artifact

A decision list, in the conversation or appended to the spec:

DECIDED:
- Cancel wins over a concurrent fire; user sees "canceled" not "published".
- Posts >2h late fail loudly rather than publish. Retry is manual.
DEFERRED:
- Rescheduling UI (owner: next iteration; cancel+recreate is the workaround).

Terms coined during the session (state names, verbs) go into CONTEXT.md while they're fresh.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

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.