consult
Skill1 starsChanged 4 months agoMITView on GitHub
Interview a new business idea or problem in CSAC order (Clarification → Structuring → Analysis → Conclusion). Draw the user's hypotheses and insight out with questions, and save a record to cases/ at the end.
21 other projects use a version of this.
Adopted by
Most recently taken up by
- ZaxbyHub/opencode-swarmtheir own version, first seen Nov 15, 2026
- aemerson-commits/shareable-skillstheir own version, first seen Nov 13, 2026
- fjwood69/moritheir own version, first seen Nov 10, 2026
- natb1/commons.systemstheir own version, first seen Nov 1, 2026
- natb1/CanonDBtheir own version, first seen Nov 1, 2026
- alimohammadiamirhossein/deputytheir own version, first seen Nov 1, 2026
Used with
Projects with it also use discover, init and 6 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.
anombyte93Skilldiscover
Phase 1 of the prd-taskmaster pipeline: brainstorm-driven discovery. Delegates to superpowers:brainstorming in Interactive Mode (one adaptive question at a time), or self-brainstorms in Autonomous Mode when no user is present. Intercepts before the brainstorming chain hands off to writing-plans — this skill owns the exit. Extracts constraints, calibrates scale (Solo / Team / Enterprise), and advances the pipeline to GENERATE.
Also in 3 projects that use this
MicrosoftSkillinit
Also in 3 projects that use this
SentrySkillreview
Fixture review skill.
Also in 5 projects that use this
AnthropicSkillstatus
Case status summary by audience — client-facing (plain language), internal (for the professor), or court-ready (formal caption format per local rules). Same facts, different framing and depth. Use when a student needs to update the client, brief the professor, or prepare a court status report.
Also in 3 projects that use this
MicrosoftSkillcouncil
Convene the persona panel — six orthogonal review lenses — on a target. Cold independent fan-out, debate-to-consensus, synthesized verdict with recorded dissent and a roster manifest.
Also in 3 projects that use this
AWSSkillship
Finalize dev work on an issue branch — update docs, push, and open a PR.
Also in 3 projects that use this
Other versions21 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 21
- zaxbysauce/opencode-swarm★ 532same name
- ZaxbyHub/opencode-swarm★ 531same name
- ansari-project/codev★ 298same name
- cluesmith/codev★ 298same name
- raine/consult-llm-mcp★ 148same name
- fjwood69/mori★ 32same name
- jamillazarev/multica-ops★ 11same name
- Reef123/OpenVera★ 9same name
- meowdiocre/meowpi★ 3same name
- alimohammadiamirhossein/deputy★ 2same name
11 more not listed
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.
- CLAUDE.mdCLAUDE.md
- pitch-reviewSkill
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 5 are distinctive, and 3 are very specific to this project.
0
Common
Most of the 17 edited versions we read keep these.
5
Distinctive
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
If they've thrown out even a one-line idea, begin.
- Very specific: line 13
Check cases/ for a file on the same topic.
- Distinctive: line 23
Exit condition: write the re-defined problem as one paragraph, show it, get their agreement.
- Distinctive: line 31
When a fact appears, push it up with "So what?" into a conclusion-shaped hypothesis ("X is true!").
- Very specific: line 41
If they're still hypothesis-hunting, run initial-research-level searches — only to the three depth targets (industry structure / major …
- Distinctive: line 44
Keep the interview alive through the verdicts: show the evidence, then hand it back — "given this, where does the hypothesis land?" They …
- Very specific: line 55
If the session went somewhere real, offer to save it as cases/YYYY-MM-DD-topic.md (using cases/_TEMPLATE.md).
- Distinctive: line 59
Don't try to finish all four stages in one session.
“Common” is measured: kept by at least 6 in 10 of the 17 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
/consult — idea consulting by interview
Dig into the user's idea or problem using the CSAC process. This is not where you analyze for them — it's where your questions make them analyze. Apply the conversation rules from CLAUDE.md (one or two questions at a time, reflect answers back structured, hypotheses from their mouth) through every stage.
Start
- If they've thrown out even a one-line idea, begin. If not, ask for the idea (or the worry) in one line.
- Check
cases/for a file on the same topic; if it exists, read it and continue. - Open by naming the four stages you're about to walk, one line each, so they have the map.
Stage 1: Clarification — fix the real problem
Goal: turn a vague idea into a problem statement with settled scope, terms, and assumptions.
- Question order: big picture → specific context → detail → validation. What to pin down: background/context, term definitions, scope (time / geography / target), available resources & authority.
- Tone: "Whose pain, exactly, does this remove?" → "Is that 'inconvenient' a time problem or a money one?" → "If we narrow the target to X, who falls out?"
- For anything they can't answer, fix it as an assumption together: "Let's assume ___ for now, and flag it so we can overturn it later."
- Exit condition: write the re-defined problem as one paragraph, show it, get their agreement.
Stage 2: Structuring — MECE decomposition and hypotheses
Goal: break the problem down into 1–3 conclusion-shaped hypotheses.
- Let them pick the decomposition axis first: "We could split this by customer journey, or by segment — which one won't cut through the heart of it?" (If needed, recommend a lens from
concepts/05-frameworks.md.) - Have them fill the branches, and check MECE with a question: "Split this way, is there a case that falls through the gap?"
- When a fact appears, push it up with "So what?" into a conclusion-shaped hypothesis ("X is true!"). Reject question-shaped ones: "That's a question — say it as a verdict."
- Prune branches that Clarification or common sense already rules out (say the reason out loud).
- Exit condition: a prioritized list of hypotheses.
Stage 3: Analysis — Key Questions, Answer First, and verification runs
Goal: a verification design per hypothesis — and, if they want it, the verification itself.
- Per hypothesis: "What would you need to see to know this is true?" → derive the Key Question. Check it against the two tests (descends from the hypothesis / useful once answered).
- "What shape of data would that answer live in?" → fix the Answer First. Then: "If you can't find data of that shape, what happens to this hypothesis?"
- If they're still hypothesis-hunting, run initial-research-level searches — only to the three depth targets (industry structure / major players / pain points), qualitative feel over precise numbers,
filetype:pdftrick included. - Once the hypothesis–KQ–AF chain stands, offer to run the checks yourself. Per KQ, search for evidence of the AF's shape — follow news citations upstream to primary sources. If one KQ needs a broad, cross-checked answer, suggest running the
deep-researchskill on it and fold its report back into the table. - Report every result as fact → implication for the hypothesis, never a data dump. Two or three solid sources per KQ is enough here — direction over precision; exact figures still wait for deep validation.
- Keep the interview alive through the verdicts: show the evidence, then hand it back — "given this, where does the hypothesis land?" They pronounce it true / no info / false. Treat "no info" as a real verdict: a hypothesis you can't verify is moot (
concepts/04-key-questions.md). - Exit condition: a [hypothesis – KQ – AF – status] table.
Stage 4: Conclusion — synthesis, memo, next actions
- Filter the surviving KQs by the three tests (contributes to the solution / touches the core / carries an interesting insight), let findings converge upward, and have them state the conclusion first: "In one sentence, what's the conclusion so far?"
- Polish it into final form, then lay out 2–3 next actions (KQs to validate deeply, things to build, people to meet).
- If the session earned a real conclusion, offer to write it up as a one-page pyramid memo (
concepts/07-pyramid-principle.md): an SCQA opening, the recommendation on top, the surviving findings as its arguments, and a "what would change our mind" line naming the assumptions still standing. Append it to the case file.
Record
If the session went somewhere real, offer to save it as cases/YYYY-MM-DD-topic.md (using cases/_TEMPLATE.md). After saving, note that "next time, call /consult topic to continue."
Notes
- Don't try to finish all four stages in one session. Follow their pace; if it breaks off, record up to the current stage.
- On each transition, announce it — "Stage 2 done, moving to Stage 3" — so they always know where they are on the map.
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.

