Results for “accusation-audit”

Audit economic analysis outputs (fiscal briefings, macro briefings, market research, longlists, and other quantitative economic documents) against methodology standards, academic literature, and common errors. Runs structured checks across core categories including counterfactual, additionality, discounting, double counting, distributional analysis, Aqua Book RIGOUR, and Flyvbjerg-style strategic misrepresentation detection. Returns a RAG scorecard with issues ranked by severity.

Verifies that every number, comparison, and scope claim in a research paper matches raw result files, using a fresh cross-model reviewer with no prior context to prevent confirmation bias.

Audit datasets for structure, missingness, labeling, suspicious values, duplicate identifiers, and documentation readiness. Use when a researcher asks for data QA, codebook review, sanity checks, or pre-analysis cleanup guidance.

Audit a repo (or diff) for weak error handling — swallowed exceptions, bare catches, unhandled promise rejections, missing network timeouts/retries, and errors logged but not surfaced. Pure Claude Code, no MCP, no signup.