yogsoth-ai
- 1.8k skills
- 0 followers
- 512 repo stars
- yesterday last updated
- ▌ Pairwise Ranking · yogsoth-aiSelect informative comparison pairs, execute judgments, update ratings, and audit coherence until ranking resolution is sufficient.
- ▌ Research Catalog · yogsoth-aiThis is the v4 capability menu. It indexes exactly 51 tactics in 10 families and tells the caller when each tactic is useful.
- ▌ Formulate Top Goal · yogsoth-aiFormalize a research goal as an outcome, desired effect, and explicit constraints.
- ▌ Identify Obstacles · yogsoth-aiEnumerate obstacles that block a target goal/direction, with type, evidence, and blocking relation.
- ▌ Statistical Testing · yogsoth-aiApply the pre-declared statistical/estimation procedure and report effect size, uncertainty, model checks, and decision-relevant interpretation.
- ▌ Falsifiability Audit · yogsoth-aiGate a hypothesis on falsifiability, operational definition, and boundary conditions; can be entered from any hypothesis-generation path.
- ▌ Structured Consensus · yogsoth-aiMap disagreement, iterate evidence/argument refinement, and terminate with an output contract containing stable consensus, unresolved disagreements, confidence/probabilities, and threshold/stop rationale.
- ▌ Crystallize North Star · yogsoth-aiCompress the goal structure and intentionality into a specific, ambitious, achievable one-sentence research North Star.
- ▌ Assumption Stress Test 2 · yogsoth-aiSurface load-bearing assumptions, classify vulnerability, challenge or perturb them, and re-check causal logic.
- ▌ Analyze Patent Citation Network · yogsoth-aiAnalyze patent citation/family relationships to identify central prior art, technological lineage, clusters, bridge patents, and citation anomalies.
- ▌ Resolve Inventive Contradiction · yogsoth-aiTurn a design or research contradiction into targeted transformations using inventive-principle and separation reasoning, then evaluate whether the transformed configuration resolves the contradiction without creating a worse one.
- ▌ Update Confidence From Evidence · yogsoth-aiUpdate confidence in a concept, relation, claim, or model component from new supporting/contradicting evidence with explicit rationale.
- ▌ Reductio Counterexample Analysis · yogsoth-aiNegate claims, derive consequences, search contradictions/counterexamples, then refine scope rather than merely defend the original wording.
- ▌ Synthesize Meta Analytic Evidence · yogsoth-aiDesign or execute a quantitative evidence synthesis with pairwise/network/cumulative/heterogeneity/bias modes while preserving study-level quality and sensitivity logic.
- ▌ Decompose Global Sensitivity · yogsoth-aiEstimate global main effects and interactions across uncertain inputs using screening/decomposition methods such as Morris→Sobol or an equivalent supplied method.
- ▌ Establish Empirical Baseline · yogsoth-aiEstablish a fair empirical baseline by inventorying methods, extracting comparable performance, normalizing conditions/compute, checking discrepancies, and estimating progress/headroom.
- ▌ Evaluate Scenario Robustness · yogsoth-aiAggregate a candidate, strategy, or portfolio across explicit scenarios under a declared robust-decision rule such as worst-case score, minimax regret, maximin, threshold survival, or pivot-trigger analysis.
- ▌ Extract Empirical Regularity · yogsoth-aiExtract a repeatable empirical pattern from observations, state its support and exceptions, and generalize cautiously without importing an unsupported mechanism.
- ▌ Optimize Design Under Budget · yogsoth-aiGiven cost per run and resource budget, choose the most information-efficient feasible design while preserving essential validity constraints.
- ▌ Select Experimental Baseline · yogsoth-aiSelect fair SOTA/simple/oracle/control baselines with provenance and rationale matched to the experimental claim.
- ▌ Verify Evidence Independence · yogsoth-aiSeek independent evidence channels sufficient to determine whether a finding/gap is source-specific. No database/provider is named.
- ▌ Analyze Constraints Readiness · yogsoth-aiCharacterize what blocks a research target: readiness dimensions, obstacles, bottlenecks, resource gaps, causal constraints, dependencies, conflicts, removability, and mitigation pathways. Depth selects lightweight triage, readiness assessment, or causal/TOC-style constraint analysis.
- ▌ Audit Explanatory Compression · yogsoth-aiTest whether an elegant/simple explanation earns its compression by forbidding alternatives, subsuming independent facts, or making risky predictions, rather than merely relabeling observations with a compact vocabulary.
- ▌ Collect Independent Judgments · yogsoth-aiCollect judgments/rankings from multiple genuinely independent evaluators or perspectives before exposing any peer outputs, preserving ballots and rationale for later social-choice aggregation.
- ▌ Estimate Performance Headroom · yogsoth-aiEstimate practical/human/theoretical ceilings and remaining performance headroom under stated assumptions.
- ▌ Formulate Meta Analysis Scope · yogsoth-aiInstantiate a typed evidence question schema such as PICO/PECO for quantitative synthesis.
- ▌ Generate Candidate Directions · yogsoth-aiGenerate diverse candidate research fields/subfields consistent with current context while permitting deliberate boundary-crossing.
- ▌ Generate Competing Hypotheses · yogsoth-aiGenerate mechanism-distinct explanations rather than superficial paraphrases.
- ▌ Identify Load Bearing Factors · yogsoth-aiIdentify factors/assumptions/uncertainties whose change most strongly controls the conclusion.
- ▌ Specify Execution Environment · yogsoth-aiSpecify environment variables that materially affect interpretation or reproducibility: hardware, software, data, configuration, and versioning.
- ▌ Audit Convergence Independence · yogsoth-aiAudit claims of independent convergence by tracing shared priors, data, framing, models, prompts, assumptions, or upstream evidence; estimate an effective independent evidence count rather than treating nominal N paths as independent.
- ▌ Calibrate Probability Forecast · yogsoth-aiAggregate and iteratively recalibrate probabilistic forecasts, preserving uncertainty, calibration evidence, and unresolved forecaster disagreement.
- ▌ Counterfactual Causal Analysis · yogsoth-aiIntervene on factors, construct counterfactual worlds, test necessity/sufficiency, and identify load-bearing causal factors.
- ▌ Detect Performance Discrepancy · yogsoth-aiDetect material score discrepancies for the same method/task across sources and propose likely explanatory condition differences.
- ▌ Synthesize Literature Evidence · yogsoth-aiBuild a reproducible literature evidence base. Scoping, systematic, deep, narrative, and snowball are execution modes selected from intent and budget, not separate strategy nodes.
- ▌ Generate Alternative Model · yogsoth-aiGenerate a plausible model variant by relaxing, replacing, or generalizing assumptions.
- ▌ Mutate Solution Population · yogsoth-aiGenerate controlled variants of a solution population through mutation/recombination operators while retaining lineage and changed dimensions.
- ▌ Normalize Comparison Scale · yogsoth-aiNormalize heterogeneous values onto a declared comparison scale while preserving direction, units, uncertainty, and the transformation rule. The normalized object may be criterion scores, resource budgets, or another comparable quantity.
- ▌ Screen Evidence Multistage · yogsoth-aiApply explicit staged inclusion/exclusion screening (title/abstract/full evidence or equivalent) with reasons and auditability.
- ▌ Select Inventive Principle · yogsoth-aiSelect one or more inventive principles appropriate to a contradiction or transformation objective and state why each principle applies.
- ▌ Update Cumulative Evidence · yogsoth-aiUpdate a cumulative evidence estimate/order as new studies or evidence blocks arrive and detect when inference changes materially over time.
- ▌ Validate Axis Independence · yogsoth-aiTest candidate dimensions pairwise for independence/redundancy; merge covarying axes and remove non-informative axes.
- ▌ Analyze Temporal Trajectory · yogsoth-aiAnalyze an ordered temporal trajectory, including persistence/narrowing/widening tests, regime changes, inflection points, and optional extrapolation with uncertainty. Entity type, window, and forecast mode are parameters.
- ▌ Assess Prior Art And Claims · yogsoth-aiAssess novelty/prior art and claim scope by decomposing claims, tracing relevant families, and checking status/provenance.
- ▌ Attach Evidence To Relation · yogsoth-aiAttach supporting or contradicting evidence to a specific claim/relation with quality, directness, and independence metadata.
- ▌ Audit Structure Consistency · yogsoth-aiDetect cycles, contradictions, orphans, type violations, and structural incoherence in a knowledge model.
- ▌ Construct Validity Envelope · yogsoth-aiSynthesize multi-axis breakpoints into a multidimensional statement of where a claim/method remains valid.
- ▌ Coverage White Space Search · yogsoth-aiMap known methods against problem/parameter space and generate candidates from uncovered and weakly covered intersections. Acquisition mechanism is runtime-selected.
- ▌ Decompose Evaluation Metric · yogsoth-aiDecompose an evaluation metric into rewarded signals, aggregation choices, polarity, ceiling effects, and Goodhart vulnerabilities.
- ▌ Decompose Research Question · yogsoth-aiDecompose a complex RQ into MECE subquestions, map dependencies, and schedule the answering order.
- ▌ Formulate Research Question · yogsoth-aiConvert a hypothesis into a precise research question with fit-for-purpose framework, scope, feasibility, and success criteria.
- ▌ Measure Portfolio Diversity · yogsoth-aiQuantify portfolio diversity/coverage across declared feature, mechanism, domain, or failure-mode dimensions and expose concentration/niche collapse.
- ▌ Trace Citation Neighborhood · yogsoth-aiExpand an evidence set through backward/forward citation chaining and terminate using explicit relevance/saturation criteria, independent of the retrieval tool used.
- ▌ Analyze Leaderboard Dynamics · yogsoth-aiAnalyze leaderboard score distributions, compression/saturation, anomalies, selective reporting, and regime shifts.
- ▌ Audit Structural Equivalence · yogsoth-aiAudit a claimed isomorphism/equivalence by constructing an explicit mapping, testing which structures/operations/invariants are preserved, searching counterexamples, and downgrading the claim to the strongest defensible relation when full isomorphism fails.
- ▌ Audit Validator Independence · yogsoth-aiAudit whether a validator, benchmark, sandbox, or test can pass by construction because it embeds the same assumptions as the theory it claims to validate. Build a claim×validator-assumption matrix and demand at least one non-circular failure channel.
- ▌ Falsification First Audit · yogsoth-aiConvert a sharp scientific claim into the cheapest decisive falsification program. Steelman the claim into a precise testable form, specify observations/computations that could break it, execute the most informative probes first, and return only BROKEN, CORROBORATED, or UNFALSIFIABLE rather than a resilience score.
- ▌ Identify Scenario Drivers · yogsoth-aiIdentify high-impact/high-uncertainty drivers and rank them for scenario construction.
- ▌ Inventory Reference Items · yogsoth-aiBuild a structured inventory of externally established reference items under a parent-specified schema, preserving applicability/status, key attributes, limitations, and provenance. Typical schemas include benchmark or known-solution/method family.
- ▌ Select Statistical Method · yogsoth-aiSelect statistical inference/estimation method from design, distribution, sample size, pairing, multiplicity, and decision objective.
- ▌ Sharpen Falsifiable Claim · yogsoth-aiSteelman a claim into its strongest precise, bounded, testable form while making hidden quantifiers, scope, mechanisms, and commitments explicit.
- ▌ Structural Transformation · yogsoth-aiDecompose a system and apply component/function transformations such as remove, substitute, combine, redistribute, divide, or reverse.
- ▌ Synthesize Field Panorama · yogsoth-aiSummarize candidate fields by maturity, competition, entry barrier, tractability, and publication/opportunity structure.
- ▌ Analyze Experiment Results · yogsoth-aiInterpret completed experimental outputs after host/runtime execution using pre-declared statistical tests, effect/uncertainty estimates, reproducibility checks, and calibrated synthesis.
- ▌ Apply Separation Principle · yogsoth-aiResolve a physical contradiction by separating conflicting requirements in time, space, scale, condition, or system level.
- ▌ Assess Evidence Saturation · yogsoth-aiEstimate marginal information gain of new evidence batches relative to the current corpus and classify continuing/near-saturation/saturated.
- ▌ Assess Patent Legal Status · yogsoth-aiClassify observed patent/legal status and flag uncertainty or jurisdiction dependence.
- ▌ Assess Ranking Consistency · yogsoth-aiAudit pairwise/ranking coherence and identify inconsistent cycles/triads.
- ▌ Assess Readiness Dimension · yogsoth-aiAssess one readiness dimension against explicit anchors/evidence.
- ▌ Build Current Reality Tree · yogsoth-aiBuild a sufficient-cause Current Reality Tree linking undesirable effects to a small root set.
- ▌ Construct Evidence Network · yogsoth-aiConstruct a network-meta-analysis evidence graph and evaluate connectivity/transitivity prerequisites.
- ▌ Define Analysis Dimensions · yogsoth-aiIdentify approximately independent dimensions/axes that span the relevant problem, design, or validity space and define their semantics.
- ▌ Design Randomness Protocol · yogsoth-aiIdentify randomness sources and define seed/repetition/propagation rules sufficient for the intended reproducibility level.
- ▌ Discover Biological Analog · yogsoth-aiIdentify biological systems that achieve the target function under comparable constraints, without assuming surface similarity.
- ▌ Enumerate Dimension Values · yogsoth-aiEnumerate meaningful values/levels for a dimension or factor. Mode controls representative, experimental, boundary, pathological, or adversarial coverage.
- ▌ Evolve Solution Population · yogsoth-aiMaintain a diverse population of candidate solutions, mutate/recombine variants, select under explicit fitness and novelty criteria, and iterate while preserving useful niches rather than collapsing prematurely to one winner.
- ▌ Extract Structural Mapping · yogsoth-aiConstruct an explicit source↔target mapping over objects, relations, operations, constraints, and invariants for a claimed structural equivalence.
- ▌ Analyze Future Scenarios · yogsoth-aiConstruct and compare plausible future/competitive/temporal/stress scenarios and test whether the research path remains robust across them.
- ▌ Apply Question Framework · yogsoth-aiSelect and instantiate a research-question schema appropriate to study type.
- ▌ Audit Benchmark Validity · yogsoth-aiTreat benchmarks as scientific measurement instruments: audit construct validity, contamination, metric pathology, coverage, leaderboard dynamics, and protocol drift.
- ▌ Audit Data Contamination · yogsoth-aiAudit train/test leakage, memorization artifacts, temporal leakage, and disclosure pathways that can invalidate evaluation.
- ▌ Construct Counterfactual · yogsoth-aiConstruct the minimally or explicitly intervened counterfactual world required by the parent analysis.
- ▌ Define Evidence Protocol · yogsoth-aiDefine research questions, evidence scope, queries/concepts, inclusion/exclusion rules, provenance requirements, and stopping evidence for an acquisition task.
- ▌ Evaluate Scenario Impact · yogsoth-aiEvaluate a fixed candidate, research path, or portfolio under one explicit scenario using stable criteria and return impact, tradeoffs, and failure triggers.
- ▌ Extract Causal Structure · yogsoth-aiExtract causal structure from source material at requested granularity: individual causal claims, ordered mechanism chains, mediators, assumptions, and stated boundary conditions.
- ▌ List Undesirable Effects · yogsoth-aiList observable undesirable effects with evidence and severity as starting points for constraint/root-cause analysis.
- ▌ Operationalize Construct · yogsoth-aiTurn abstract variables/constructs into measurable operational definitions.
- ▌ Optimize Pareto Frontier · yogsoth-aiGenerate a non-dominated frontier for a multi-objective candidate/portfolio problem.
- ▌ Predict Competitive Move · yogsoth-aiPredict plausible competitor research moves, timing, and preemption/priority risk with explicit assumptions.
- ▌ Probe Benchmark Artifact · yogsoth-aiProbe whether benchmark performance can be explained by shortcuts, artifacts, leakage, annotation cues, metric gaming, or non-target capabilities.
- ▌ Represent Mechanism Edge · yogsoth-aiRepresent a directed mechanism edge with pathway, sign/direction, assumptions, falsifier, strength, and evidence provenance.
- ▌ Select Solution Variants · yogsoth-aiSelect candidate variants under fitness, novelty, constraint, and diversity criteria while avoiding premature population collapse.
- ▌ Trace Assumption Cascade · yogsoth-aiTrace downstream claims, methods, and conclusions that fail when a selected assumption is invalidated; distinguish local from cascading failure.
- ▌ Assess Construct Validity · yogsoth-aiAssess whether a benchmark measures its claimed capability using content, convergent, discriminant, and confound analysis.
- ▌ Assess Patent Claim Scope · yogsoth-aiAssess protection breadth, limiting elements, support, overlap, design-around vulnerability, and likely scope uncertainty for a patent claim without making legal conclusions beyond the evidence.
- ▌ Assess Problem Wickedness · yogsoth-aiAssess whether the problem is tame, complex, or wicked by examining stakeholder disagreement, moving boundaries, feedback, value conflict, and absence of a stable stopping rule.
- ▌ Construct Perspective Set · yogsoth-aiConstruct an explicit set of scientifically relevant viewpoints, roles, worldviews, stakeholder positions, or epistemic lenses to apply to the same object.
- ▌ Design Falsification Test · yogsoth-aiSpecify a concrete observation, computation, intervention, counterexample, or measurement that would count against the sharpened claim, prioritizing cheap high-information tests.
- ▌ Explore Dimensional Space · yogsoth-aiRepresent a problem/research space as validated dimensions and values, enumerate compatible combinations, and optionally map occupancy, detect gaps, prioritize regions, or synthesize candidate questions/ideas.
- ▌ Map Stakeholder System · yogsoth-aiModel the research system from stakeholder and boundary perspectives: define what is inside/outside the system, identify stakeholder jobs and stakes, classify salience, and expose conflicts that alter problem framing or feasibility.