The question itself may be the bug. Use when a debate goes in circles, both options of a binary feel wrong, improving one parameter degrades another, "we're stuck" on a problem that resists direct attack, a metaphor is doing hidden work in the architecture, or a trade-off is being denied rather than made.
Scanned 8/31/2026
Install to Claude Code
npx -y skills add cdeust/zetetic-team-subagents --skill problem-reframing --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Problem Reframing?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/cdeust-problem-reframing)More formats (shields.io, HTML) on the badges page.
---
name: problem-reframing
shapes: [problem-reframing]
description: >
The question itself may be the bug. Use when a debate goes in circles, both
options of a binary feel wrong, improving one parameter degrades another,
"we're stuck" on a problem that resists direct attack, a metaphor is doing
hidden work in the architecture, or a trade-off is being denied rather than made.
---
# Problem Reframing
**Problem shape:** effort is being spent on a question that is malformed —
a false binary, a contradiction treated as a compromise, a metaphor mistaken
for the mechanism, a dispute about words masquerading as a dispute about facts.
The move: dissolve, restate, or invert the frame before solving anything.
## Relevant geniuses
| Agent | Use when |
|---|---|
| [wittgenstein](../../agents/genius/wittgenstein.md) | the dispute is about how words are being used — audit the language game; some problems dissolve rather than resolve |
| [nagarjuna](../../agents/genius/nagarjuna.md) | both sides of a binary feel wrong — run the tetralemma (true / false / both / neither) |
| [aristotle](../../agents/genius/aristotle.md) | "why" questions conflated — separate the four causes; catalog the fallacies in play |
| [midgley](../../agents/genius/midgley.md) | a metaphor (machine, war, market, brain) silently shapes the design — audit where it breaks down |
| [leguin](../../agents/genius/leguin.md) | a trade-off is being denied — force the genuine cost onto the table; audit the narrative frame of the plan |
| [eco](../../agents/genius/eco.md) | the artifact will be interpreted — design for the model reader; bound the open vs closed interpretations |
| [bateson](../../agents/genius/bateson.md) | the conflict escalates symmetrically or the instructions double-bind — the pathology is in the interaction, not the parties |
| [altshuller](../../agents/genius/altshuller.md) | improving one parameter degrades another — formulate the contradiction and resolve it, don't compromise |
| [polya](../../agents/genius/polya.md) | stuck on a problem that resists direct attack — restate, work backward, find the related solved problem, specialize |
| [schon](../../agents/genius/schon.md) | the current approach keeps failing mid-action — surface the implicit frame and switch strategy deliberately |
## Invocation
1. Pick the best-fit agent above. If two or more fit, run
`tools/genius-invoker.sh route "<problem>"` and take the top ranked match.
2. Load it: `tools/genius-invoker.sh invoke <agent> "<problem>"`, then read
`agents/genius/<agent>.md` in full.
3. Apply the agent's `<workflow>` step by step and answer in its
`<output-format>`. The deliverable is the reframed question plus what the
new frame makes visible — then route the reframed problem to the right
skill (estimation, causal-audit, formal-correctness, ...).
4. Typical chain: wittgenstein dissolves the verbal dispute → altshuller
formulates the residual contradiction → polya attacks it. Run via
`tools/genius-invoker.sh compose wittgenstein altshuller -- "<problem>"`.
5. If no shape above matches, use a standard team agent instead.
## Refuse when
- The frame is fine and the problem is just hard — reframing as procrastination.
- The requester wants the reframe to make a real trade-off disappear.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!