FORMA capabilities are grouped by the constraints they solve in real operation. Legacy module names (MFCore, Vault, Pulse, Immunity, Formatr, Mollee) map into these capability surfaces rather than standing alone as disconnected products.
Each capability is evaluated by the same criteria: does it reduce execution risk before runtime, does it improve operator judgment quality, and does it preserve provenance for audit and reproducibility. This keeps the system coherent as features evolve.
What it does: Schedules and executes governed runs on Material Cloud execution infrastructure.
Why it exists: Material execution requires profile-bound scheduling, not ad-hoc runtime intervention.
Artifacts produced: Produces run records with profile context, execution metadata, and outcome linkage.
What it does: Authoring layer for writing .matr programs, compiling Molebytes, versioning artifacts, local simulation, and governance checks.
Why it exists: Expression quality determines downstream reliability and governance readiness.
Artifacts produced: Produces source revisions, compile outputs, and authoring provenance.
What it does: Models Matter inside Material constraints and validates behavior before execution requests are submitted.
Why it exists: Preflight modeling in Material Twin reduces high-cost mistakes and improves confidence at scale.
Artifacts produced: Produces Material Twin traces, baseline outputs, and validation evidence.
What it does: Maintains immutable artifact and run lineage as a system of record.
Why it exists: Governance depends on attributable history and reproducible execution context.
Artifacts produced: Produces version graphs, run lineage, and provenance chains.
What it does: Captures structured outcomes, confidence indicators, and comparative diffs.
Why it exists: Material execution decisions require signal interpretation rather than log tailing.
Artifacts produced: Produces signal cards, variance reports, and decision-ready summaries.
What it does: Applies risk classes, cost ceilings, approvals, and execution boundaries.
Why it exists: Agent-native operation is only acceptable when boundaries are explicit and enforced.
Artifacts produced: Produces policy decisions, denials, approvals, and audit records.
What it does: Provides operator surfaces and agent orchestration loops under the same governance model.
Why it exists: Adoption requires automation with accountability, not hidden autonomous pathways.
Artifacts produced: Produces human review actions, route decisions, and escalation outcomes.
Every capability must consume or produce artifacts that originate from Matter (.matr) source workflows.
Execution pathways remain bounded by explicit profiles, cost ceilings, approvals, and escalation ladders.
Capabilities should map to Material Twin evidence and stated non-claims in the Research corpus.
