Agent Access Receipt Lab v0.1.6: Synthetic Data and Code for Agent-Mediated Content/API Access Decisions

This record contains a fully synthetic data and code package for evaluating agent-mediated content/API access decisions as auditable receipt records. The package includes 15 scenario cases, 15 mutation/control cases, 3 receipt-integrity controls, generated synthetic receipts, a JSON Schema, CSV/JSON result tables, and a self-contained Python reproduction script. Version v0.1.6 preserves the scenario, mutation/control, integrity-control, required-field/schema-shape guard, and verifier outcomes while using neutral synthetic labels and abstract resource identifiers throughout the package. It also updates the public DOI/version metadata for this Zenodo version. The corpus uses no human-subject data, real traffic logs, private publisher telemetry, operational API logs, production credentials, or third-party confidential material. The HMAC tokens are public deterministic fixtures for reproducibility only and are not a production security mechanism.本数据集包含一套全合成数据与代码包，用于评估由AI智能体（AI Agent）介导的内容/应用程序编程接口（Application Programming Interface，API）访问决策，并以可审计回执记录（auditable receipt records）的形式留存评估相关数据。该代码包涵盖15组场景用例、15组变异/对照用例、3组回执完整性校验控件、合成生成的回执文件、JSON模式（JSON Schema）、CSV与JSON格式结果表格，以及一套自包含的Python复现脚本。 v0.1.6版本保留了场景用例、变异/对照用例、完整性校验控件、必填字段/模式形状校验规则与验证器结果，同时在整个代码包中统一使用中性合成标签与抽象资源标识符；该版本还更新了本Zenodo存档的公开数字对象标识符（Digital Object Identifier，DOI）与版本元数据。 本数据集未包含人类受试者数据、真实流量日志、发布方私有遥测数据、运营API日志、生产环境凭据或第三方机密资料。哈希消息认证码（Hash-based Message Authentication Code，HMAC）令牌仅为用于保障可复现性的公开确定性固定值，并非生产级安全机制。