Crisistrace
An incident and investigation workspace that links instrumented agent decisions, approvals, wallet execution, transaction hashes, multi-hop fund-flow hypotheses, and reviewer conclusions without claiming identity or causality from chain data alone.
Existing blockchain forensics products trace fund flows, while agent runtimes can record prompts, data, tool calls, approvals, and wallet execution. The unserved join is valuable only when instrumented before the incident: a later model log and an on-chain transaction do not prove one caused the other. Crisistrace should bind a signed decision and authorization identifier into the transaction execution record, preserve both raw chains of evidence, and present multi-hop attribution as reviewable hypotheses rather than evidence-grade fact or confirmed counterparty identity.
The compliance, security, investigation, or operations leader at a digital-asset firm running instrumented agents that can prepare or execute on-chain transactions.
Current agent trading and regulatory attention make the gap timely.
Compliance, security, and investigations teams have a concrete agent-transaction incident job.
The input has limited direct graph support.
A clear digital-asset investigation buyer, enterprise forensics validation, available indexing and cross-chain primitives, and an unserved agent-execution join support the direction.
The graph support is weak, instrumentation must predate the incident, on-chain attribution is probabilistic, India alignment is not established, and enterprise incumbents can add agent telemetry.
Discussion
No comments yet — be the first to weigh in.