Screen one record's organisation for adverse indicators, and say what was not screened.
Maps to a partner-backed resource type and runs against one saved record at a time, on demand. Screens that record's organisation for sanctions, PEP, adverse media, regulatory enforcement, insolvency, disqualification, litigation and corporate actions via the Ploutonion organization-compliance agent — all ten IndicatorType values every time, so a type no source can answer comes back as a coverage gap rather than a silent omission. The agent's response is stored verbatim under adverse-indicators-checks/ and an entry is appended to adverse-indicators-checks.json beside the record: the raw file is what the agent said, the log is what we concluded about it, and the log points at the findings rather than copying them. Screens the named entity only — parent companies and beneficial owners are not screened, and every outcome and every log entry says so. The organisation name must come from identity.json (a name typed on the record is not identification), so an unidentified record is skipped with a reason and nothing is written for it. The same skill bundle carries the analyst's verdict entry point: the screen can never return a FalsePositive, so clearing a finding is a human decision, and a later check suppresses a cleared finding and records that it did. Never invents a finding. Does not run ISO or modern-slavery discovery.