chore: scaffold PYRE MVP monorepo (structure + docs)

pnpm + TypeScript workspace per design doc §13:
- apps/{web,api,worker} skeletons (Next.js 16, Fastify 5, BullMQ)
- packages/{core,solana,prometheus,db,config} (core has real types/DTOs;
  solana/prometheus are stubs)
- programs/pyre-core placeholder (future Anchor, v1.0)
- docs/: PYRE_MVP_DESIGN (canonical), ARCHITECTURE, SECURITY, TOKEN_CLASSIFICATION
- CLAUDE.md, README, .env.example (no private-key var by design)

Skeleton + docs only — no Solana/business logic yet. All workspaces typecheck clean.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

packages/prometheus/src/index.ts

@@ -0,0 +1,72 @@
+/**
+ * @pyre/prometheus — AI generation logic (STUBS ONLY).
+ *
+ * Responsibilities (§13): prompt templates, the meta mixer, output parser,
+ * safety checks, and the image-prompt generator. See §9 (Prometheus AI Meta
+ * Mixer) and §10 (Pump.fun Creator Workflow) of `docs/PYRE_MVP_DESIGN.md`.
+ *
+ * Design notes:
+ * - Prometheus generates Spawn *identity* only — it never controls funds.
+ * - Meta influence is probabilistic, not deterministic.
+ * - Do not allow users to force exact copyrighted/existing meme identities;
+ *   produce inspired mutations, not direct clones.
+ *
+ * TODO: the AI client (Anthropic / OpenAI / image-gen provider) is configured
+ * via `@pyre/config` and injected here — this package adds NO SDK dependencies
+ * of its own. Wire the client through function parameters or a small interface.
+ *
+ * Nothing here is implemented yet — every function throws "not implemented".
+ */
+import type { PrometheusInput, PrometheusOutput } from "@pyre/core";
+
+const NOT_IMPLEMENTED = "not implemented";
+
+/**
+ * Build the text-generation prompt from mixer input.
+ *
+ * TODO: apply prompt templates and the probabilistic meta-influence weighting
+ * (burned archetypes / Essence themes / chaos mutation / operator seed).
+ */
+export function buildPrompt(_input: PrometheusInput): string {
+  throw new Error(NOT_IMPLEMENTED);
+}
+
+/**
+ * Run the meta mixer end-to-end: build prompt, call the AI client, parse, and
+ * run safety checks, producing a Spawn package.
+ *
+ * TODO: orchestrate buildPrompt -> AI client -> parseOutput -> runSafetyChecks.
+ * The AI client is injected (configured via @pyre/config).
+ */
+export function runMetaMixer(_input: PrometheusInput): Promise<PrometheusOutput> {
+  throw new Error(NOT_IMPLEMENTED);
+}
+
+/**
+ * Parse a raw model response into a structured `PrometheusOutput`.
+ *
+ * TODO: validate/normalize the model JSON into the output schema.
+ */
+export function parseOutput(_raw: string): PrometheusOutput {
+  throw new Error(NOT_IMPLEMENTED);
+}
+
+/**
+ * Run safety/compliance checks over generated output.
+ *
+ * TODO: moderation + copyright/clone guards; return risk flags. Must reject
+ * attempts to clone exact copyrighted or existing meme identities.
+ */
+export function runSafetyChecks(_output: PrometheusOutput): Promise<string[]> {
+  throw new Error(NOT_IMPLEMENTED);
+}
+
+/**
+ * Generate an image prompt for the Spawn artwork.
+ *
+ * TODO: derive an image prompt from the Spawn identity, honoring the same
+ * anti-clone safety rules.
+ */
+export function generateImagePrompt(_output: PrometheusOutput): string {
+  throw new Error(NOT_IMPLEMENTED);
+}