When Claims Become Auditable: Distributed AASC/UEAP Claim-Status Audit and the Game Theory of Scientific Authority

Scientific gatekeeping has traditionally concentrated the power to validate claims in journals, referees, editorial boards, professional societies, prestige networks, and institutional review processes. This manuscript analyzes a different strategic condition: claim-status audit becomes portable, partially mechanizable, and non-permissioned. The central question is not whether journals adopt a new internal review procedure, but what happens when institutions can no longer prevent structured claim-status audit from being run externally. The Admissibility and Standing Constraint framework (AASC), operationalized through the Universal Exactness Audit Protocol (UEAP), supplies the audit grammar. A capable language model, human evaluator, or hybrid system supplies an execution layer; it does not author the protocol’s target, carrier, blocker, status, laundering, severity, or reportability categories. When the protocol and a target manuscript are placed in the same evaluation context, the process can produce a registry-style classification of claim target, requested report status, carrier, scope, calibration ancestry, unresolved blockers, possible report-status laundering, severity, and strongest reportable formulation. The paper distinguishes four layers that distributed-audit rhetoric often collapses: audit execution, audit publication, audit standing, and audit consequence. Execution can be open without every output becoming authoritative. Audit standing is earned through target fidelity, basis declaration, execution provenance, passage grounding, registry completeness, reproducibility, severity discipline, counteraudit, and right of reply. Consequence is a further social or institutional response that should be proportional to the audit’s standing and severity. Distributed audit therefore democratizes execution, not authority. The game-theoretic model separates raw audit occurrence from standing-bearing consequential detection. If potential auditor k initiates an audit with probability pk, detects the relevant mismatch with probability qk, reaches an uptake-sensitive audience with probability rk, and produces an audit with standing weight wk, then δw eff = 1 − Y k (1 − pkqkrkwk) is the benchmark probability of at least one standing-bearing consequential detection under independent source events. The paired benchmark ηw eff = 1 − Y k (1 − pkfkrkwk) tracks standing-weighted false-positive exposure, where fk is the conditional probability of a false or materially overstated finding. The standing term is an audit-output property, not an auditor-prestige score, and any implementation of wk = W(Pk,Rk,Ck, Sk,Hk) is constrained by declared provenance, replication, convergence, severity, response-handling, and consequencespecific admissibility rules. The central design question is therefore whether standing-bearing correction value exceeds standing-bearing false-finding harm, audit noise, chilling effects, and capture risk. The analysis develops volunteer audit supply, self-audit, institutional audit reversal, external venue classification, pressure migration to audit-basis manipulation, symbolic compliance, model and prompt variance, audit-basis fragmentation, prompt injection, harassment, severity governance, and registry memory. Institutions do not disappear, and LLM-assisted audits do not become truth oracles. Rather, institutions lose exclusive control over whether structured, protocolized claim-status audit can be executed against accessible claims and become p