question-forge
Skill4 starsChanged 48 days agoMITView on GitHub
Take a surface question and forge the question actually worth asking — diagnose what's broken in the asked question (smuggled answers, wrong level, false binary, self-protection), transform it, and return a better question WITHOUT answering it. Use when a question feels stuck, loaded, or too easy, or as a preprocessor before other reasoning skills.
In this repository48 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .agents/skills/AGENTS.mdAGENTS.md
- .agents/skills/assumption-audit/AGENTS.mdAGENTS.md
- .agents/skills/decision-matrix/AGENTS.mdAGENTS.md
- .agents/skills/direct-setup/AGENTS.mdAGENTS.md
- .agents/skills/direct-verify/AGENTS.mdAGENTS.md
- .agents/skills/first-principles-thinking/AGENTS.mdAGENTS.md
- .agents/skills/percolate-kb/AGENTS.mdAGENTS.md
- .agents/skills/percolate-reading/AGENTS.mdAGENTS.md
- .agents/skills/phase-final-reviewer/AGENTS.mdAGENTS.md
- .agents/skills/phase-implementer/AGENTS.mdAGENTS.md
- .agents/skills/phase-orchestrator/AGENTS.mdAGENTS.md
- .agents/skills/phase-reviewer/AGENTS.mdAGENTS.md
- .agents/skills/plan-kb/AGENTS.mdAGENTS.md
- .agents/skills/pr-cleanup/AGENTS.mdAGENTS.md
- .agents/skills/query-kb/AGENTS.mdAGENTS.md
- .agents/skills/question-forge/AGENTS.mdAGENTS.md
- .agents/skills/reclaim-disk-space/AGENTS.mdAGENTS.md
- .agents/skills/refresh-kb/AGENTS.mdAGENTS.md
- .agents/skills/riff/AGENTS.mdAGENTS.md
- .agents/skills/save-pdf-kb/AGENTS.mdAGENTS.md
- .agents/skills/save-url-kb/AGENTS.mdAGENTS.md
- .agents/skills/write-phase-plan/AGENTS.mdAGENTS.md
- docs/AGENTS.mdAGENTS.md
- kb/AGENTS.mdAGENTS.md
- personal-website/AGENTS.mdAGENTS.md
- scripts/AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- assumption-auditSkill
- decision-matrixSkill
- direct-setupSkill
- direct-verifySkill
- first-principles-thinkingSkill
- percolate-kbSkill
- percolate-readingSkill
- phase-final-reviewerSkill
- phase-implementerSkill
- phase-orchestratorSkill
- phase-reviewerSkill
- plan-kbSkill
- pr-cleanupSkill
- query-kbSkill
- reclaim-disk-spaceSkill
- refresh-kbSkill
- riffSkill
- save-pdf-kbSkill
- save-url-kbSkill
- write-phase-planSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈2
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 18
If the question is genuinely operational — factual, technical, well-posed ("why is the build failing?", "what's the capital gains rate?") — …
- Distinctive: line 78
Never answer the forged question: If the user wants it answered, they can ask — or pipe it into another skill.
≈ “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
Question Forge
Use this skill to reformulate a question rather than answer it. The premise, borrowed from fiction: a great novel doesn't answer "is revenge justified?" — it constructs the version of that question that becomes concrete, costly, and unavoidable. The illumination is in the formulation. Accordingly, this skill's contract inverts the usual one: the output is a question, not an answer.
Purpose
Most questions arrive broken in predictable ways. Answering a broken question efficiently just delivers the asker to the wrong destination faster. This skill diagnoses the break, applies transformations, tests candidates against the standards a dramatist would use, and hands back the forged question — explicitly refusing to answer it.
Procedure
1. Dispatch check (mandatory, first)
If the question is genuinely operational — factual, technical, well-posed ("why is the build failing?", "what's the capital gains rate?") — say the question is fine and answer it normally. Being preciously Socratic about factual questions is this skill's primary failure mode.
2. Diagnose the asked question
Common breaks:
- Smuggled answer — the question presupposes its conclusion ("isn't it time we...")
- Wrong level — asks tactics when the problem is values, or values when the problem is tactics
- False binary — two options presented where the space is wider
- Displaced subject — asks about the world when it's really about the asker ("should I quit?" is rarely about the job)
- Wish in question syntax — a complaint or longing wearing a question mark
- Shield question — a safe question standing guard in front of a harder one nearby
3. Generate reformulations along transformation axes
- Subject shift — world → self: "should I quit?" → "what am I hoping quitting will fix?"
- Level shift — up or down the ladder of abstraction
- Time shift — this decision → the shape of a life; prospective → retrospective ("what will I wish I had asked?")
- Inversion — "why can't I X?" → "what is not-X-ing accomplishing for me?"
- Value excavation — "what would have to be true for this to matter?"
- Dramatization — embody it in a concrete scene with stakes: not "is loyalty owed to institutions?" but "the whistleblower's sister works on the same floor"
4. Test candidates against the dramatist's criteria
A forged question should:
- Sustain a novel — resist one-line resolution
- Have live opposition — opposing answers both carry real weight; if one side is obviously right it's a quiz, not a question
- Implicate the asker — not stay safely third-person
- Cost something — answering it honestly requires giving something up
- Stay generative — keep producing thought after you stop looking at it
5. Deliver — and hold the line
Present the forged question. Do not answer it.
Output format
## The question you asked
...
## What it's doing
<the diagnosis: what the question smuggles, protects, or displaces>
## The forged question
**...**
## What changed
<which transformation(s), and why this version has more torque>
## Living with it
<one short paragraph: what it would look like to carry this question for a while, rather than resolve it>Guardrails
- One pass. Forge once; do not reformulate the reformulation. No infinite regress.
- Livable, not a koan. The forged question must be one a person can actually carry and act under — not mystical wordplay.
- The original deserves respect. Diagnose without condescension; the asked question is the raw material, not an error.
- Never answer the forged question. If the user wants it answered, they can ask — or pipe it into another skill.
Composition
Works as a preprocessor for the rest of this package: forge the question, then feed the result to dp-solve, decision-matrix, assumption-audit, or parable.
Canonical questions
- "Should I quit my job?"
- "Is AI going to take our jobs?"
- "Was I right to end the relationship?"
- "What should our company's mission be?"
- "Why is engagement dropping?" (often forges into: "what did we build this for?")
Poor fits: operational questions (the dispatch check handles these), and questions the user has already carefully formulated — sometimes the question arrives correct.
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.

