Narrative surface

Agent runtime coordination

Disclaimer: this page is non-enforcing and replay-safe. It does not confer autonomous execution authority, orchestration privilege, or protected payload reconstruction. All linked demos remain read-only / verifier-derived.

Multi-agent work splits across workers while a coordinator holds selection context steady. Continuity overlays echo what replay already captured — aligning with Verifier-safe exchange and portable bundles without claiming live cross-runtime enforcement.

Coordinator (read-only)Agent AAgent BAgent CTrust continuity overlay (replay-safe projections — non-enforcing)

Trust continuity overlays

  • Runtime projections stay tenant-scoped; overlays summarize digest-aligned state without mutating ontology.
  • Bounded replay snapshots anchor what the overlay can cite — derived-only narration.

Helper lineage · GhostLog continuity

  • Helper and GhostLog-derived lines show step lineage auditors can reconcile in Replay Studio.
  • No execution claims: continuity is explanatory copy over replay-visible structure.

Recovery pivots · bounded autonomy

  • Recovery pivots appear where replay narratives classify transitions — governance remains external to this page.
  • Autonomy is bounded by policy surfaces elsewhere; this route makes no remediation authority statement.

Runtime trust handshakes · verifier continuity

  • Handshake and interoperability envelopes (LP-009 / LP-010) describe digest echoes — inspect via Replay Studio export panels.
  • Verifier continuity stays JSON-local; paste flows use the same carriesExecutionAuthority: false posture as other verifier UX.