agentleFS
TheGoat395TheGoat395/Codex-Skills/skills/creativity/SKILL.md

creativity

Skill136 starsChanged 5 months agoMITView on GitHub

Generate original ideas or escape fixation.

4 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 10 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈10

    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
  1. Distinctive:

    Create possibilities that are meaningfully different, useful for the objective, and honest about evidence.

    line 8
  2. Distinctive:

    Do not confuse creativity with random novelty, many synonyms, ornamental effects, unsupported claims, or a long unranked list.

    line 10
  3. Distinctive:

    Comparing or selecting candidates: read convergence protocol and selection and validation.

    line 17
  4. Distinctive:

    Invent: create a mechanism, format, or concept not implied by the current defaults.

    line 29
  5. Distinctive:

    Write a compact challenge statement:

    line 46
  6. Distinctive:

    If the initial framing presupposes one solution, restate it in functional terms before generating ideas.

    line 54
  7. Distinctive:

    use multiple structurally different references rather than one dominant exemplar.

    line 62
  8. Distinctive:

    test whether the final idea would still exist if the exemplar disappeared.

    line 66
  9. Distinctive:

    If the user explicitly requests faithful adaptation or reconstruction, preserve that objective and distinguish required fidelity from …

    line 68
  10. Distinctive:

    Stop when the search has covered genuinely different mechanisms, the leading direction survives the material constraints, further variants …

    line 72

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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.

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

Creativity

Create possibilities that are meaningfully different, useful for the objective, and honest about evidence. Creativity is a search process: frame the problem, generate varied preinventive material, explore it from several interpretations, develop the strongest candidates, and evaluate only after genuine alternatives exist.

Do not confuse creativity with random novelty, many synonyms, ornamental effects, unsupported claims, or a long unranked list.

Route generation and convergence

For a brief reframe, explore distinct mechanisms and develop the result directly. Respect an already selected direction.

  • Substantial ideation, synthesis, invention, repair or search-heavy planning: read concept development and creative search methods; use generation/exploration, flexibility/persistence and distinct families.
  • Comparing or selecting candidates: read convergence protocol and selection and validation; separate generation/ranking, preserve fatal constraints and test decisive uncertainty.

critical-thinking assesses claims; logical-thinking determines what follows. Imagined audience reactions are hypotheses, not validation.

Choose the creative task

Identify what kind of change is needed:

  • Reframe: expose a more productive question, tension, audience, or unit of value.
  • Diverge: produce distinct solution families before committing.
  • Synthesize: combine distant but compatible mechanisms into a coherent whole.
  • Transform: change the representation, sequence, scale, medium, or interaction model.
  • Invent: create a mechanism, format, or concept not implied by the current defaults.
  • Repair: escape a stale, generic, overconstrained, or repeatedly failing direction.

If installed, creative-direction-divergence can own a formal website direction study and cross-domain-creative-divergence a business or campaign brief. Neither is required by this public collection. Otherwise use the search, development and selection references below, adapting evaluation criteria to the brief. Preserve an already selected direction unless revisiting it is in scope.

Frame a useful search space

Record:

  1. desired outcome and the human or system it serves;
  2. the central tension, unmet need, or failure in the obvious answer;
  3. fixed constraints versus inherited conventions that may be challenged;
  4. the ordinary baseline and the dimensions on which an idea must differ;
  5. available assets, evidence, mechanisms, and implementation capacity;
  6. what success means beyond novelty;
  7. the smallest uncertainty worth exploring first.

Write a compact challenge statement:

Create [outcome] for [audience/context] while honoring [fixed constraints],
without defaulting to [obvious pattern]. A strong direction changes [decisive dimension]
and can be tested by [observable signal].

If the initial framing presupposes one solution, restate it in functional terms before generating ideas.

Control fixation before exposure

Examples can improve fluency while anchoring output to their surface features. When originality matters:

  • define functions, constraints, and evaluation dimensions before inspecting examples;
  • describe a reference by the mechanism that works, not by its visible styling or wording;
  • use multiple structurally different references rather than one dominant exemplar;
  • include at least one deliberately distant source domain;
  • state which surface features must not be copied;
  • generate one candidate before exposure when the task permits;
  • test whether the final idea would still exist if the exemplar disappeared.

If the user explicitly requests faithful adaptation or reconstruction, preserve that objective and distinguish required fidelity from creative freedom.

Stop and deliver

Stop when the search has covered genuinely different mechanisms, the leading direction survives the material constraints, further variants repeat existing families, and the next uncertainty is better resolved by a prototype or evidence than more ideation.

Match the output to the user's need:

  • for exploration, present a curated set of distinct developed directions;
  • for a decision, lead with the selected direction, why it wins, its sacrifice, and its smallest test;
  • for implementation, translate the concept into concrete requirements while preserving its organizing mechanism;
  • for an unresolved search, name the competing directions and the observation that would discriminate them.

Label invention, analogy, assumption, inspiration, and verified fact separately. Do not expose private scratch reasoning; show the useful creative structure and result.

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.

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.