Searchable workflow library
Find copy-ready prompts for Codex, ChatGPT, Claude, Gemini, and practical AI workflows. Filter by outcome, category, tool, difficulty, or keyword.
Document AI inputs, accepted and rejected outputs, source checks, student revisions and learner-supplied learning evidence without certifying authorship or institutional compliance.
Convert assigned sources into a claim-to-evidence ledger that separates source findings, student interpretation, counterevidence and unresolved uncertainty before drafting.
Inspect a research package for data provenance, licensing, environment, code, seeds, dependencies and expected outputs, separating reproduced evidence from inaccessible work.
Build and stress-test a qualitative codebook using traceable excerpts, rival interpretations, coder disagreement, positionality and documented adjudication.
Cross-check hypotheses, outcomes, exclusions, power assumptions, models, multiplicity controls and decision rules before data inspection, with deviations made explicit.
Test whether a research question can be answered with the proposed design, population, variables, data and estimand, separating repairable gaps from fatal limitations.
Decide whether preliminary research is mature enough to inform a policy or operating decision, with bounded claims, reversible use, and monitoring conditions.
Determine whether cited evidence remains current and authoritative after corrections, retractions, updates, superseding guidance, or changed decision conditions.
Test whether a consequential analytical conclusion survives plausible changes to data, cohort, definitions, assumptions, model choices, and missing-information treatment.
Reconstruct where a data product diverged from authoritative lineage, bound affected outputs and decisions, and define safe repair and reprocessing.
Challenge forecast assumptions, structural stability, backtest evidence, scenario sensitivity, and decision thresholds before relying on projected outcomes.
Tune agent escalation triggers using incident severity, uncertainty, false-positive and false-negative evidence, queue capacity, delay, and owner authority.