Convert theory-first research ideas into proof-oriented analysis plans.
npx skills add https://github.com/shenmuxing/EtaSkill --skill analysis-plan-shenmuxing
Please help me install this Agent Skill. Skill: analysis-plan Source: https://github.com/shenmuxing/EtaSkill/tree/main/skill-examples/analysis-plan Command: npx skills add https://github.com/shenmuxing/EtaSkill --skill analysis-plan-shenmuxing
This Skill helps convert an early-stage theory idea into a compact, proof-oriented research plan so you can clarify what must be proved, what assumptions are needed, and where the idea could fail before investing in experiments.
Use the analysis-plan skill to turn this research idea into a theorem target, list the assumptions, and draft the proof obligations.
High-intent search queries and answers about installing and using this skill.
A proof plan maps your assumptions to specific claims, identifies proof levers, and outlines disproof paths so you can check if a theorem is provable or falsifiable before investing in experiments.
Proof obligation mapping decomposes a theory-first hypothesis into precise claims, supporting lemmas, and independent failure modes to create an auditable, publication-ready research plan.
Assess disproof paths by surfacing boundary cases and mapping failure modes to determine whether your hypothesis is falsifiable, provable, or too weak to pursue before committing to a full proof.
Yes, the approach supports convergence analysis and lower-bound framing by requiring explicit assumptions, claim decomposition, and boundary-case validation to structure these non-experimental research workflows.
Validate boundary cases and assumptions when converting an early-stage theory idea into an auditable research plan to clarify what must be proved and identify where the idea could fail before experimentation.
Theory-first planning is limited to non-experimental workflows and may not suit hypotheses that cannot be framed as theorems, counterexamples, or convergence claims, requiring explicit assumptions and claim decomposition to proceed.
shenmuxingGenerate and rank theory-first research ideas from papers or open problems.
THUFanZdConstruct and verify mathematical proofs for ML/AI theory.
jkfeeCross-check proposed method claims against recent literature for novelty.
Weiwei-MaoGenerate and test LLM-driven hypotheses from tabular datasets.
