§02f Jupiter — Post-Hoc Forensics — What this layer covers
Mars® Spec › §02f Jupiter — Post-Hoc Forensics › What this layer covers
Section index · The evidence trail →
Mars® Protocol — §02f: Post-Hoc Forensic Layer
Version 1.0-draft
1. What this layer covers
This layer specifies the post-hoc forensic layer of the Mars® governed invocation: the structures, interfaces, and composition discipline governing the analysis of a retained evidence trail after the Mars® conformance certificate has been issued.
The post-hoc forensic layer operates on the evidence trail and its retained witnesses. It does not re-perform the conformance verification that produced the certificate; it operates on the verified record that verification produced. Its outputs — attributed findings, divergence signals, posture states, longitudinal verdicts, reinforcement actions — are forensic artifacts that are themselves order-typed, evidenced, and recomputable against the same witness infrastructure that governs the certificate.
What this layer does not cover. This layer does not specify the conformance certificate (§02b), the boundary crossing (§02c), the analysis battery (§02a), or the pre-invocation interface (§02e). References to those layers are cross-references to their governing sections, not re-specifications. Artifact use — knowledge bases, data sources — in the context of analysis is governed by §03.
Dependencies. This layer depends on §02a (order-decomposition, multi-stage witness chain, composition discipline, analytical battery), §02b (conformance certificate, recomputation witness, delta-attestation lifecycle, specification registry), §02c (actuation anchor, boundary artifact classes), §03-iii §2.12 (attribution — the two-layer KB × data-source attribution join is the governed substrate for the post-hoc artifact attribution mechanism of §8; the fallback hierarchy and join result record specified there are the same artifact-relation machinery used here in the post-hoc context; an implementation of §8 that does not ground post-hoc artifact attribution in the §03-iii §2.12 join is not practicing the governed attribution mechanism), and §03-iii §2.8 (symmetric reinforcement — the reinforcement-action layer of §7 is downstream of the symmetric reinforcement signals; a reinforcement action routed without consuming the §03-iii §2.8 typed reinforcement signal under the declared reinforcement-policy record is not a governed reinforcement action under this layer). Post-hoc forensic artifacts produced by this layer may feed back into subsequent governed invocations through the actuation anchor and the pre-invocation signal registry (§02e §3), but the feedback path is a hand-off at the boundary — not a loop within this layer.