rethink
Skill0 starsChanged 5 months agoMITView on GitHub
Use when you have an existing solution, design, draft, plan, or code structure and need genuinely different approaches — not tweaks or variations. Triggers: feeling anchored, "try another way", "rethink this", a solution that works but feels uninspired, or an approach that keeps hitting the same wall.
5 other projects use a version of this.
Adopted by
Most recently taken up by
- epicenter-md/epicentertheir own version, first seen Oct 8, 2026
- EpicenterHQ/epicentertheir own version, first seen Oct 8, 2026
- qiulin/skillstheir own version, first seen Sep 10, 2026
- gyujeongion/claude-code-rootcausetheir own version, first seen Jun 27, 2026
- fabioespindula/awesome-nanoclaw-skillstheir own version, first seen May 7, 2026
Used with
Projects with it also use adversarial-review.
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.
MicrosoftSkilladversarial-review
Adversarial review of a system design from 6 critical perspectives -- SRE, security, staff engineer, finance, operator, and developer advocate. Produces a unified risk assessment. Use for INTERACTIVE on-demand reviews during a design conversation (/adversarial-review). For RECIPE-DRIVEN reviews (where prior step context is needed), use the systems-design-critic agent instead.
Also in 3 projects that use this
Other versions5 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 5
- epicenter-md/epicenter★ 4.8ksame name
- EpicenterHQ/epicenter★ 4.8ksame name
- fabioespindula/awesome-nanoclaw-skills★ 2same name
- gyujeongion/claude-code-rootcause★ 1same name
- qiulin/skills★ 0same name
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.
- lean-searchSkill
- youtube-learnSkill
The file
We split this skill into the rules it gives an agent
2 rules. 0 are common, 1 is distinctive, and 1 is very specific to this project.
0
Common
Most of the 4 edited versions we read keep these.
1
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 18
NOT for greenfield ideation (use brainstorming).
- Distinctive: line 27
If you can't name the assumption a candidate drops, it's a variation — discard it.
“Common” is measured: kept by at least 6 in 10 of the 4 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
Rethink
Overview
A different approach must abandon a load-bearing assumption of the current one. Anything that keeps all the original's assumptions is a variation, not an alternative. This skill is for escaping a frame you are already inside.
NOT for greenfield ideation (use brainstorming). NOT for polishing a solution
already verified correct (don't churn working code).
The Iron Law
NO ALTERNATIVE COUNTS UNLESS IT DROPS A LOAD-BEARING ASSUMPTION OF THE ORIGINAL.If you can't name the assumption a candidate drops, it's a variation — discard it.
The Process
1. Strip to the goal
State what success means in one sentence, with zero mention of how the current solution does it. ("The user needs to ___.") This is the target every alternative must hit; the current solution is just one path to it.
2. Expose the current frame
List the load-bearing assumptions the current approach bakes in — the choices so ingrained they feel like the problem itself (e.g. "data is relational", "the user reads top-to-bottom", "work happens on request", "it's a form"). For each, ask: what if this weren't true?
3. Diverge — apply at least 3 operators, each yielding a different family
- Invert — reverse a core assumption.
- Drop the constraint — what if the hardest limit simply vanished?
- Change the primitive — a different fundamental building block (form→conversation, table→graph, request/response→stream, page→canvas).
- Borrow a domain — how would a game designer / a spreadsheet / biology / a physical store solve this?
- Go to extremes — the radically minimal version, and the radically maximal version.
- Move the work — shift who/when/where it happens (user↔system, client↔server, build-time↔runtime, sync↔async).
4. Distinctness gate
For each candidate, name the assumption it drops. Two candidates that drop the same assumption and look alike are one idea wearing two hats — merge them or push one further. Aim for 2–4 survivors that are genuinely different in kind.
5. Contrast and recommend
For each survivor: what it makes easy that the original made hard, and what it sacrifices. Then recommend one (or a synthesis) tied to the actual goal and constraints — explicitly say why it beats the original, or honestly conclude the original wins after a real fight.
Red Flags — rationalizations to reject
| Thought | Reality |
| "These options are all distinct." | If they drop the same assumption, it's one idea in three outfits. Name the dropped assumption for each. |
| "The original is basically fine." | Maybe — but defending it before diverging is anchoring. Earn that verdict in step 5, not step 0. |
| "I already considered alternatives." | Considering ≠ generating. Actually run the operators. |
| "A fresh approach is riskier." | The gate is distinctness, not adoption. Generate freely; judge risk in step 5. |
| Changed colors / params / wording / file layout. | Surface change. Did the organizing principle move? If not, keep going. |
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.