agent-teams-command
Use when work has genuinely independent streams or distinct builder, evaluator, domain, and integration roles that require…
Use when a defined problem needs diverse ideas, cross-domain combinations, and cheap experiments instead of a single untested answer.
$ npx -y skills add Mark393295827/third-brain-v7-skills --skill creativity-engine --agent claude-codeHow it fires
How this skill gets triggered: by you, by Claude, or both.
/creativity-engineContext preview
The summary Claude sees to decide when to auto-load this skill.
Use when a defined problem needs diverse ideas, cross-domain combinations, and cheap experiments instead of a single untested answer.
name: creativity-engine description: Use when a defined problem needs diverse ideas, cross-domain combinations, and cheap experiments instead of a single untested answer. metadata: version: "8.1.0" updated: "2026-08-18" profile: "one-shot" assumes: "A problem, target user, and at least one meaningful constraint can be stated." conflicts_with: "Unbounded brainstorming, novelty without utility, or treating generated ideas as validated opportunities."
<skill_contract> <input>Defined problem, target user, desired change, constraints, prior attempts, and experiment budget.</input> <output>Three non-equivalent options with scored mechanisms and falsifiable minimum experiments.</output> <done>Each selected option has a hypothesis, smallest artifact, participant, threshold, budget, stop rule, and failure learning.</done> <non_goals>Unbounded brainstorming, novelty without utility, polished prose before diversity, or presenting ideas as validated demand.</non_goals>
Create option value by combining mechanisms, constraints, and analogies, then convert the strongest options into minimum experiments.
Provide: problem, target user, desired change, constraints, existing attempts, and experiment budget. Optional: domains or concepts to combine.
<intake>
Rewrite the request as `For [user], change [state] under [constraints], measured by [signal]`. Extract reusable building blocks: actors, assets, mechanisms, channels, incentives, and constraints.
</intake>
<unknowns_gate>
If no user, problem, or constraint is available, return `NEEDS_INPUT` with one discriminating probe. Treat market demand, technical feasibility, and user behavior as testable unknowns, not assumptions to hide.
</unknowns_gate>
<execute>
1. Generate 10-20 combinations across at least three mechanisms or domains. 2. Include inversion, subtraction, constraint removal, and one distant analogy. 3. Cluster duplicates by underlying mechanism, not wording. 4. Score survivors on expected value, test difficulty, distinctiveness, reversibility, and evidence gap. 5. Select three non-equivalent options. 6. For each, define a minimum experiment: hypothesis, smallest artifact, target participant, success threshold, budget, stop rule, and learning captured on failure.
Do not optimize prose before option diversity. A useful failed experiment is better than an impressive concept with no falsifier.
</execute>
<evaluate>
Check that the top three differ in mechanism, fit constraints, expose their largest unknown, and can be tested cheaply. Remove ideas that are only features, slogans, or unsupported scale claims. Recombine once if all finalists share the same failure mode.
</evaluate>
Return `status`, `result` (idea clusters and top-three experiments), `evidence` (inputs and scoring basis), `unknowns`, and `next_action`.
</skill_contract>
Verification-first Agent Skills for Codex OS, with optional compatibility adapters for Claude Code, Gemini, Cursor, and Windsurf.
Use when work has genuinely independent streams or distinct builder, evaluator, domain, and integration roles that require…
Use when designing or refactoring a model-native engineering workflow with bounded autonomy, probes, custom evaluation, durable…
Use when property-service operations need an AI plus ontology plus DMAIC design for work orders, dispatch, quotes, evidence, CTQ…
Use when a personal or team operating system needs a bounded redesign using Four-C, closed-loop controls, 70/30 allocation, 3B…
Use when a goal must be converted into a repeatable behavior, cue, SOP, review cadence, and identity-aligned reinforcement.
Use when source material must be transformed into a compact, evidence-aware model for learning, decisions, or an Obsidian concept…