Stats
Actions
Tags
From pratyabhijna-creative-engine
Use for scientific or mathematical creativity tasks - hypothesis generation, analogy-finding across domains, BIG-Bench-Hard "Causal Judgement"-style problems, or any "give me a non-obvious explanation/analogy" prompt. The cascade is configured for higher cit-temperature and a wider sampler grid to favor exploration.
How this skill is triggered — by the user, by Claude, or both
Slash command
/pratyabhijna-creative-engine:scientific-creativityThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Hypothesis brainstorming, cross-domain analogy hunting, "non-obvious explanation" tasks, BIG-Bench-Hard creative reasoning, or whenever the user wants "the kind of leap that wins a paper, not the standard answer".
Hypothesis brainstorming, cross-domain analogy hunting, "non-obvious explanation" tasks, BIG-Bench-Hard creative reasoning, or whenever the user wants "the kind of leap that wins a paper, not the standard answer".
constraint_text = "non-obvious explanation/analogy for: {question}".must_avoid to standard textbook explanations the user has already heard.aspects from 2-3 candidate framings (e.g. ["thermodynamic", "graph-theoretic", "biological"]).pratyabhijna_mcp__cascade with:
K=8max_tokens=128render_mode="verbatim"vimarsa_event = true, the cascade found a cross-frame analogy; surface it with the diagnostics.delta_F and the runner-up.Hypothesis: <the BMR winner>
Cross-frame analogy: <if vimarsa fired, name which frames>
Why this is non-obvious: <novelty score, contrast with retrieval set>
Runner-up: <next-best>
λ_a = 2.0, λ_p = 2.0. Increase λ_p if must_avoid is doing the heavy lifting.top_p = 0.90 and temperature ≤ 1.0 (the LM shouldn't hallucinate axioms).Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub sharathsphd/pratyabhijna --plugin pratyabhijna-creative-engine