Spectre.Execution.Materialization (Spectre v0.3.0)

Exact, verified boundary between Skill routing and data-driven execution.

The struct carries the closed program, resolved input and typed prompt plans needed by the shared operational executor. Its portable identity contains only digests and lineage, so logs and projections do not expose prompt or input content.

Summary

Types

t()

Functions

to_data(materialization)

Returns the privacy-safe portable identity of a materialization.

verify(materialization)

Verifies every internal digest and exact plan/receipt binding.

Types

t()

@type t() :: %Spectre.Execution.Materialization{
  continuation_id: String.t(),
  definition_ref: String.t(),
  digest: String.t(),
  input: term(),
  mount_id: term(),
  plans: %{optional(String.t()) => Spectre.Prompt.Plan.t()},
  program: Spectre.Execution.Program.t(),
  projection: Spectre.Projection.t(),
  prompt_receipts: [Spectre.Prompt.Receipt.t()],
  route_label: term(),
  schema_version: pos_integer()
}

Functions

to_data(materialization)

@spec to_data(t()) :: map()

Returns the privacy-safe portable identity of a materialization.

verify(materialization)

@spec verify(t()) :: :ok | {:error, term()}

Verifies every internal digest and exact plan/receipt binding.