@manifesto-ai/studio-core

Projection-first read-only analysis engine for Manifesto domains.

Overview

@manifesto-ai/studio-core takes a DomainSchema plus optional overlays and returns stable JSON projections for tools and dashboards.

Use it when you want:

a semantic graph of the domain
findings analysis over static and runtime overlays
action availability and coarse blocker explanations, alongside runtime-owned fine intent-legality tooling
snapshot, trace, lineage, and governance inspection
one read-only analysis engine shared across CLI, MCP, and UI surfaces

createStudioSession(bundle, options?) is the supported entry point.

Entry Point

typescript

import { createManifesto } from "@manifesto-ai/sdk";
import { createStudioSession } from "@manifesto-ai/studio-core";

const runtime = createManifesto(schema, effects).activate();

const session = createStudioSession({
  schema,
  snapshot: runtime.inspect.canonicalSnapshot(),
  trace,
  lineage,
  governance,
}, {
  validationMode: "lenient",
});

Session Methods

Method	Purpose
`attachSnapshot(snapshot)`	Attach a canonical runtime snapshot
`attachTrace(trace)`	Attach a `TraceGraph`
`attachLineage(lineage)`	Attach lineage export or query-like input
`attachGovernance(governance)`	Attach governance export or query-like input
`detachOverlay(kind)`	Drop one overlay from the session
`getGraph(format?)`	Return the domain graph projection
`getFindings(filter?)`	Return findings across the current bundle
`getActionAvailability()`	Return coarse action-family legality from the current canonical snapshot
`explainActionBlocker(actionId)`	Explain why an action family is blocked
`inspectSnapshot()`	Return a snapshot inspection projection
`analyzeTrace()`	Return trace replay analysis
`getLineageState()`	Return lineage state projection
`getGovernanceState()`	Return governance state projection

Projection Contracts

API	Projection
`getGraph()`	`DomainGraphProjection`
`getFindings()`	`FindingsReportProjection`
`getActionAvailability()`	`ActionAvailabilityProjection[]`
`explainActionBlocker()`	`ActionBlockerProjection`
`inspectSnapshot()`	`SnapshotInspectorProjection`
`analyzeTrace()`	`TraceReplayProjection`
`getLineageState()`	`LineageStateProjection`
`getGovernanceState()`	`GovernanceStateProjection`

All of these are JSON-serializable and surface-neutral.

Input Rules

snapshot must be a canonical runtime snapshot from runtime.inspect.canonicalSnapshot()
lineage and governance accept canonical exports and widened query-like shapes
absent optional overlays degrade to structured "not-provided" results instead of failing the whole session
malformed optional overlays throw only when validationMode: "strict" is enabled

getGraph() and getFindings() always work with schema alone.

Where It Fits

@manifesto-ai/studio-cli wraps this package for terminal workflows.
@manifesto-ai/studio-mcp wraps the same analysis surface as MCP tools.
UI packages and dashboards should build on this package instead of re-implementing graph or findings logic.
Runtime Tooling Surface maps Studio-style tooling to the public runtime, snapshot, and lineage seams.

@manifesto-ai/studio-core ​

Overview ​

Entry Point ​

Session Methods ​

Projection Contracts ​

Input Rules ​

Where It Fits ​

Related Docs ​