Research

Present the GoalTree to the user for confirmation. Ask about reasonableness, missing elements, and priority ordering among sub-goals.

Understand hard boundaries on the user's research — target venues, methodology preferences, areas to avoid, advisor/team requirements. Not limited to ML/AI — works for any research domain.

Detect annotation artifacts and shortcuts in benchmarks

Extract and steel-man the core arguments supporting a given opinion cluster.

Distill the strongest arguments from each perspective through Argument Delphi or Dialectical Delphi methods.

Establish acceptability standards through RAND/UCLA Appropriateness Method or Consensus Conference protocols.

Find positive deviants and reframe the problem from deficit-based to asset-based using Appreciative Inquiry.

Search for positive deviants and extract transferable principles using Appreciative Inquiry.

Challenge industry best practices' hidden assumptions. Deconstruct benchmarks to reveal unexamined constraints.

SOP: 设计子问题的最优回答顺序

Tactic: 归纳/溯因路径——描述异常现象，生成候选解释，按可信度排序

SOP: 描述和分类无法被现有理论解释的异常现象

KAOS-style recursive goal decomposition. AND decomposition for sub-goals that must ALL be satisfied. OR decomposition for alternative paths where any one suffices. Produces a GoalTree (DAG structure).

Assess analogy depth (surface/structural/systemic). Determines whether an analogy warrants transfer investment.

Extract transferable structural principles from source domains. Orchestrates source identification → abstraction → structural mapping → transfer validation.

Chain analogies to deeper levels (3-5 layers). Each layer reveals new aspects and insights not visible at the surface.

Systematic structure-mapping from source to target domain (Gentner). Identify relational correspondences and transfer higher-order constraints.

Generate alternatives for every known approach — ensure no approach goes unchallenged.

**Date:** 2026-02-02 20:47:34 **Owner:** Claude Code (Research) **Phase:** Research **Git Commit:** 4fa9eb15d09e1263d0bd443f4ac7469b52d9904b **Last Updated:** 2026-02-02 20:47:34 **Last Updated By:** Claude Code (Research) **Tags:** [performance, bottlenecks, agent, tools, streaming]

Score each candidate alternative against all criteria to produce a score matrix.

Generate alternative model formulations by relaxing, replacing, or generalizing specific assumptions.

Generate 2-4 divergent scenarios from the same evidence base, each representing a plausible alternative to the artifact's conclusions.

Strategy: What-If Analysis, Alternative Futures, and Four Ways of Seeing — generate competing explanations and scenarios to challenge the dominant narrative.

Classify gap root causes using AHRQ 4-reason framework (insufficient info, biased info, inconsistent info, not yet integrated).