Case Compendium 50-Case Public Release Candidate: Source-Bounded Failure Corpus, Methodology Paper, and First Algorithmic Public Administration Synthesis

This release contains the first public release candidate of the Case Compendium, a 50-case source-bounded corpus of high-hazard accidents, public administration failures, medical and public-health governance failures, environmental exposure cases, labor and household burden dockets, market-failure cases, benchmark-governance failures, and consumer-protection failures. The corpus snapshot is frozen against: Case Compendium Master _CS001-0x-0021_ This public package includes a curated/redacted workbook, a methodology paper, a first bounded synthesis report, a maintenance and impact plan, source-link tracking materials, release notes, citation metadata, and a CC BY 4.0 license notice. The central purpose of the compendium is methodological: it preserves case boundaries, source anchors, rival explanations, evidence tiers, mechanism candidates, remedy paths, and public-claim limits so that later synthesis can be performed against a frozen record rather than through memory, anecdote, or premature unification. The first synthesis report focuses on an algorithmic and automation-adjacent public administration cluster, including Robodebt, Post Office Horizon, Dutch childcare benefits, Windrush, COMPAS, AFST / child welfare risk scoring, the UK 2020 exam algorithm, and Michigan MiDAS. The provisional candidate pattern is that a proxy, score, system record, or administrative rule can become action-bearing while affected people carry the burden of correction through weak, delayed, opaque, or underpowered review pathways. This release does not claim that the corpus is externally validated, statistically representative, predictive, or complete. All case classifications and mechanism calls remain author-coded / reviewer-pending unless and until independent coding is recorded in a later release. The package is intended for inspection, review, teaching, correction, citation, and bounded synthesis. AI assistance was used for organization, drafting, formatting, and workbook preparation. AI is not treated as evidentiary authority. Case selection, source interpretation, claim boundaries, classifications, and public responsibility remain with the author/curator. This public release is licensed under Creative Commons Attribution 4.0 International (CC BY 4.0), except where otherwise noted. Third-party sources cited or linked in the workbook remain under their original rights and licenses. Keywords case compendium; institutional failure; public administration; algorithmic governance; automated decision systems; risk governance; source-bounded corpus; failure analysis; public accountability; reviewer-pending dataset; methodology; synthesis firewall; high-hazard systems; environmental justice; consumer protection; public health governance; CC BY 4.0