@forwardimpact/libsyntheticprose 0.1.1

package/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to the Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2026 Dick Olsson
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

package/engine/pathway.js

@@ -0,0 +1,342 @@
+/**
+ * Pathway Engine — orchestrates LLM calls to generate pathway entity data.
+ *
+ * Generates entities in dependency order:
+ * framework → levels → stages → behaviours → capabilities →
+ * drivers → disciplines → tracks → self-assessments
+ *
+ * @module libuniverse/engine/pathway
+ */
+
+import { readFileSync } from "fs";
+import { join } from "path";
+import { buildFrameworkPrompt } from "../prompts/pathway/framework.js";
+import { buildLevelPrompt } from "../prompts/pathway/level.js";
+import { buildStagePrompt } from "../prompts/pathway/stage.js";
+import { buildBehaviourPrompt } from "../prompts/pathway/behaviour.js";
+import { buildCapabilityPrompt } from "../prompts/pathway/capability.js";
+import { buildDriverPrompt } from "../prompts/pathway/driver.js";
+import { buildDisciplinePrompt } from "../prompts/pathway/discipline.js";
+import { buildTrackPrompt } from "../prompts/pathway/track.js";
+
+/**
+ * Load JSON schemas from the schema directory.
+ * @param {string} schemaDir - Path to products/map/schema/json/
+ * @returns {object} schemas keyed by entity type
+ */
+export function loadSchemas(schemaDir) {
+  const names = [
+    "framework",
+    "levels",
+    "stages",
+    "behaviour",
+    "capability",
+    "discipline",
+    "track",
+    "drivers",
+    "self-assessments",
+    "defs",
+  ];
+  const schemas = {};
+  for (const name of names) {
+    schemas[name] = JSON.parse(
+      readFileSync(join(schemaDir, `${name}.schema.json`), "utf-8"),
+    );
+  }
+  return schemas;
+}
+
+/**
+ * PathwayGenerator orchestrates LLM calls to generate pathway entity data.
+ */
+export class PathwayGenerator {
+  /**
+   * @param {import('./prose.js').ProseEngine} proseEngine - Prose engine for LLM calls
+   * @param {object} logger - Logger instance
+   */
+  constructor(proseEngine, logger) {
+    if (!proseEngine) throw new Error("proseEngine is required");
+    if (!logger) throw new Error("logger is required");
+    this.proseEngine = proseEngine;
+    this.logger = logger;
+  }
+
+  /**
+   * Generate all pathway entity data via LLM calls in dependency order.
+   * @param {object} options
+   * @param {object} options.framework - Framework AST from DSL parser
+   * @param {string} options.domain - Universe domain
+   * @param {string} options.industry - Universe industry
+   * @param {object} options.schemas - Loaded JSON schemas
+   * @returns {Promise<object>} Generated pathway data keyed by entity type
+   */
+  async generate({ framework, domain, industry, schemas }) {
+    return generatePathwayData({
+      framework,
+      domain,
+      industry,
+      schemas,
+      proseEngine: this.proseEngine,
+    });
+  }
+}
+
+/**
+ * Generate all pathway entity data via LLM calls in dependency order.
+ *
+ * @param {object} options
+ * @param {object} options.framework - Framework AST from DSL parser
+ * @param {string} options.domain - Universe domain
+ * @param {string} options.industry - Universe industry
+ * @param {object} options.schemas - Loaded JSON schemas
+ * @param {import('./prose.js').ProseEngine} options.proseEngine - Prose engine for LLM calls
+ * @returns {Promise<object>} Generated pathway data keyed by entity type
+ */
+async function generatePathwayData({
+  framework,
+  domain,
+  industry,
+  schemas,
+  proseEngine,
+}) {
+  const ctx = { domain, industry };
+
+  // 1. Framework metadata
+  const fw = await generateEntity(
+    "framework",
+    "framework",
+    buildFrameworkPrompt(framework, ctx, schemas.framework),
+    proseEngine,
+  );
+
+  // 2. Levels
+  const levels = await generateEntity(
+    "levels",
+    "levels",
+    buildLevelPrompt(framework.levels, ctx, schemas.levels),
+    proseEngine,
+  );
+
+  // 3. Stages
+  const stages = await generateEntity(
+    "stages",
+    "stages",
+    buildStagePrompt(framework.stages, ctx, schemas.stages),
+    proseEngine,
+  );
+
+  // 4. Behaviours (parallel — no cross-references)
+  const behaviours = await Promise.all(
+    framework.behaviours.map((b) =>
+      generateEntity(
+        "behaviour",
+        b.id,
+        buildBehaviourPrompt(b, ctx, schemas.behaviour),
+        proseEngine,
+      ).then((data) => ({ ...data, _id: b.id })),
+    ),
+  );
+
+  // 5. Capabilities with skills (parallel)
+  const capabilities = await Promise.all(
+    framework.capabilities.map((c, i) =>
+      generateEntity(
+        "capability",
+        c.id,
+        buildCapabilityPrompt(
+          { ...c, ordinalRank: i + 1 },
+          ctx,
+          schemas.capability,
+        ),
+        proseEngine,
+      ).then((data) => ({ ...data, _id: c.id })),
+    ),
+  );
+
+  // Collect all skill IDs and behaviour IDs from DSL declarations
+  // (not from LLM output — these must be available even in no-prose mode)
+  const skillIds = framework.capabilities.flatMap((c) => c.skills || []);
+  const behaviourIds = framework.behaviours.map((b) => b.id);
+
+  // 6. Drivers (reference skills + behaviours)
+  const drivers = await generateEntity(
+    "drivers",
+    "drivers",
+    buildDriverPrompt(
+      framework.drivers,
+      { ...ctx, skillIds, behaviourIds },
+      schemas.drivers,
+    ),
+    proseEngine,
+  );
+
+  // 7. Disciplines (reference skills, behaviours, track IDs from DSL)
+  const trackIds = framework.tracks.map((t) => t.id);
+  const disciplines = await Promise.all(
+    framework.disciplines.map((d) =>
+      generateEntity(
+        "discipline",
+        d.id,
+        buildDisciplinePrompt(
+          d,
+          { ...ctx, skillIds, behaviourIds, trackIds },
+          schemas.discipline,
+        ),
+        proseEngine,
+      ).then((data) => ({ ...data, _id: d.id })),
+    ),
+  );
+
+  // 8. Tracks (reference capability IDs for skillModifiers)
+  const capabilityIds = framework.capabilities.map((c) => c.id);
+  const tracks = await Promise.all(
+    framework.tracks.map((t) =>
+      generateEntity(
+        "track",
+        t.id,
+        buildTrackPrompt(
+          t,
+          { ...ctx, capabilityIds, skillIds, behaviourIds },
+          schemas.track,
+        ),
+        proseEngine,
+      ).then((data) => ({ ...data, _id: t.id })),
+    ),
+  );
+
+  // 9. Self-assessments (deterministic — no LLM)
+  const selfAssessments = generateSelfAssessments(
+    framework,
+    skillIds,
+    behaviourIds,
+  );
+
+  return {
+    framework: fw,
+    levels,
+    stages,
+    behaviours,
+    capabilities,
+    drivers,
+    disciplines,
+    tracks,
+    selfAssessments,
+  };
+}
+
+/**
+ * Generate a single entity via the prose engine.
+ *
+ * @param {string} entityType - Entity type for cache key prefix
+ * @param {string} entityId - Entity ID for cache key
+ * @param {{ system: string, user: string }} prompt - Built prompt
+ * @param {import('./prose.js').ProseEngine} proseEngine - Prose engine
+ * @returns {Promise<object|null>} Parsed JSON data
+ */
+async function generateEntity(entityType, entityId, prompt, proseEngine) {
+  const key = `pathway:${entityType}:${entityId}`;
+  const result = await proseEngine.generateJson(key, [
+    { role: "system", content: prompt.system },
+    { role: "user", content: prompt.user },
+  ]);
+  return result;
+}
+
+/**
+ * Simple seeded PRNG (mulberry32). Deterministic given the same seed.
+ * @param {number} seed
+ * @returns {() => number} Returns values in [0, 1)
+ */
+function createRng(seed) {
+  let s = seed | 0;
+  return () => {
+    s = (s + 0x6d2b79f5) | 0;
+    let t = Math.imul(s ^ (s >>> 15), 1 | s);
+    t = (t + Math.imul(t ^ (t >>> 7), 61 | t)) ^ t;
+    return ((t ^ (t >>> 14)) >>> 0) / 4294967296;
+  };
+}
+
+/**
+ * Pick a random index near a base, weighted toward ±1 with rare ±2 outliers.
+ * @param {() => number} rng - Seeded random function
+ * @param {number} base - Centre index for this level
+ * @param {number} max - Maximum valid index (inclusive)
+ * @returns {number} Clamped index
+ */
+function jitter(rng, base, max) {
+  const r = rng();
+  // 50% same, 20% +1, 15% -1, 10% +2, 5% -2
+  let offset = 0;
+  if (r < 0.5) offset = 0;
+  else if (r < 0.7) offset = 1;
+  else if (r < 0.85) offset = -1;
+  else if (r < 0.95) offset = 2;
+  else offset = -2;
+  return Math.max(0, Math.min(max, base + offset));
+}
+
+/**
+ * Generate self-assessments with realistic randomized distributions.
+ *
+ * Each assessment centres skills around the expected proficiency for
+ * that level, then applies per-skill jitter so profiles look natural:
+ * most skills cluster near the base, with occasional outliers.
+ * Behaviours use tighter jitter (±1 only, less variance).
+ *
+ * @param {object} framework - Framework AST
+ * @param {string[]} skillIds - All skill IDs from capabilities
+ * @param {string[]} behaviourIds - All behaviour IDs
+ * @returns {object[]}
+ */
+function generateSelfAssessments(framework, skillIds, behaviourIds) {
+  const proficiencies = framework.proficiencies || [
+    "awareness",
+    "foundational",
+    "working",
+    "practitioner",
+    "expert",
+  ];
+  const maturities = framework.maturities || [
+    "emerging",
+    "developing",
+    "practicing",
+    "role_modeling",
+    "exemplifying",
+  ];
+
+  const seed = framework.seed || 1;
+  const rng = createRng(seed);
+  const maxP = proficiencies.length - 1;
+  const maxM = maturities.length - 1;
+
+  const assessments = [];
+  const levelNames = ["junior", "mid", "senior", "staff", "principal"];
+
+  for (let i = 0; i < Math.min(levelNames.length, proficiencies.length); i++) {
+    const skillProficiencies = {};
+    for (const skillId of skillIds) {
+      skillProficiencies[skillId] = proficiencies[jitter(rng, i, maxP)];
+    }
+
+    const behaviourMaturities = {};
+    for (const behaviourId of behaviourIds) {
+      // Behaviours use tighter variance: ±1 only (no ±2 outliers)
+      const r = rng();
+      let offset = 0;
+      if (r < 0.55) offset = 0;
+      else if (r < 0.8) offset = 1;
+      else offset = -1;
+      behaviourMaturities[behaviourId] =
+        maturities[Math.max(0, Math.min(maxM, i + offset))];
+    }
+
+    assessments.push({
+      id: `example_${levelNames[i]}`,
+      skillProficiencies,
+      behaviourMaturities,
+    });
+  }
+
+  return assessments;
+}

package/engine/prose.js

@@ -0,0 +1,222 @@
+/**
+ * Prose Engine — LLM-assisted prose generation with cache.
+ *
+ * Uses libllm for completions, libutil for cache key hashing,
+ * and libprompt for prompt template loading.
+ */
+
+import { readFileSync, writeFileSync, existsSync } from "fs";
+import { dirname, join } from "path";
+import { fileURLToPath } from "url";
+import { generateHash } from "@forwardimpact/libutil";
+import { createLogger } from "@forwardimpact/libtelemetry";
+import { PromptLoader } from "@forwardimpact/libprompt";
+
+const __dirname = dirname(fileURLToPath(import.meta.url));
+
+export class ProseEngine {
+  /**
+   * @param {object} options
+   * @param {string} options.cachePath      Path to .prose-cache.json
+   * @param {string} options.mode           "cached" | "generate" | "no-prose"
+   * @param {boolean} [options.strict]      Fail on cache miss
+   * @param {import('@forwardimpact/libllm').LlmApi} options.llmApi
+   *        Pre-configured LLM client — required when mode is "generate"
+   * @param {import('@forwardimpact/libprompt').PromptLoader} options.promptLoader
+   *        Prompt template loader
+   * @param {object} options.logger         Logger instance
+   */
+  constructor({
+    cachePath,
+    mode,
+    strict = false,
+    llmApi,
+    promptLoader,
+    logger,
+  }) {
+    if (!cachePath) throw new Error("cachePath is required");
+    if (!mode) throw new Error("mode is required");
+    if (!promptLoader) throw new Error("promptLoader is required");
+    if (!logger) throw new Error("logger is required");
+    this.cachePath = cachePath;
+    this.mode = mode;
+    this.strict = strict;
+    this.llmApi = llmApi;
+    this.promptLoader = promptLoader;
+    this.logger = logger;
+    this.cache = this.#loadCache();
+    this.dirty = false;
+    this.stats = { hits: 0, misses: 0, generated: 0 };
+  }
+
+  /**
+   * Generate or retrieve prose for a key.
+   * @param {string} key
+   * @param {object} context
+   * @returns {Promise<string|null>}
+   */
+  async generateProse(key, context) {
+    if (this.mode === "no-prose") return null;
+
+    const cacheKey = generateHash(key, JSON.stringify(context));
+
+    if (this.cache.has(cacheKey)) {
+      this.stats.hits++;
+      return this.cache.get(cacheKey);
+    }
+
+    if (this.mode === "cached") {
+      this.stats.misses++;
+      if (this.strict) throw new Error(`Cache miss: '${key}'`);
+      return null;
+    }
+
+    // Tier 1: generate via libllm
+    const prose = await this.#callLlm(key, context);
+    this.stats.generated++;
+    if (prose) {
+      this.cache.set(cacheKey, prose);
+      this.dirty = true;
+    }
+    return prose;
+  }
+
+  /**
+   * @param {string} key
+   * @param {object} context
+   * @returns {Promise<string|null>}
+   */
+  async #callLlm(key, context) {
+    const prompt = this.#buildPrompt(key, context);
+    const response = await this.llmApi.createCompletions({
+      messages: [
+        { role: "system", content: this.promptLoader.load("prose-system") },
+        { role: "user", content: prompt },
+      ],
+      max_tokens: context.maxTokens || 500,
+    });
+    const content = response.choices?.[0]?.message?.content?.trim() || null;
+    this.logger.info("prose", `Generated: ${key}`, {
+      chars: content ? content.length : 0,
+    });
+    return content;
+  }
+
+  /**
+   * Generate or retrieve a structured response (pre-built messages).
+   * @param {string} key - Cache key
+   * @param {object[]} messages - Pre-built messages array [{role, content}]
+   * @returns {Promise<string|null>}
+   */
+  async generateStructured(key, messages) {
+    if (this.mode === "no-prose") return null;
+
+    const cacheKey = generateHash(key, JSON.stringify(messages));
+
+    if (this.cache.has(cacheKey)) {
+      this.stats.hits++;
+      this.logger.debug("prose", `Cache hit: ${key}`);
+      return this.cache.get(cacheKey);
+    }
+
+    if (this.mode === "cached") {
+      this.stats.misses++;
+      if (this.strict) throw new Error(`Cache miss: '${key}'`);
+      return null;
+    }
+
+    const response = await this.llmApi.createCompletions({
+      messages,
+      max_tokens: 4000,
+    });
+    const content = response.choices?.[0]?.message?.content?.trim() || null;
+    this.stats.generated++;
+    if (content) {
+      this.cache.set(cacheKey, content);
+      this.dirty = true;
+    }
+    this.logger.info("prose", `Generated structured: ${key}`, {
+      chars: content ? content.length : 0,
+    });
+    return content;
+  }
+
+  /**
+   * Generate structured output and parse as JSON.
+   * @param {string} key - Cache key
+   * @param {object[]} messages - Pre-built messages array
+   * @returns {Promise<object|null>}
+   */
+  async generateJson(key, messages) {
+    const raw = await this.generateStructured(key, messages);
+    if (!raw) return null;
+    const cleaned = raw
+      .replace(/^```(?:json)?\s*\n?/m, "")
+      .replace(/\n?```\s*$/m, "")
+      .trim();
+    return JSON.parse(cleaned);
+  }
+
+  /** @returns {Map<string, string>} */
+  getProseMap() {
+    return this.cache;
+  }
+
+  saveCache() {
+    if (!this.dirty || !this.cachePath) return;
+    writeFileSync(
+      this.cachePath,
+      JSON.stringify(Object.fromEntries(this.cache), null, 2),
+    );
+    this.dirty = false;
+  }
+
+  /**
+   * Build a prompt from key and context.
+   * @param {string} key
+   * @param {object} context
+   * @returns {string}
+   */
+  #buildPrompt(key, context) {
+    return this.promptLoader.render("prose-user", {
+      topic: context.topic || key.replace(/_/g, " ").replace(/-/g, " "),
+      tone: context.tone || "technical",
+      length: context.length || "2-3 paragraphs",
+      domain: context.domain,
+      role: context.role,
+      audience: context.audience,
+      scenario: context.scenario,
+      driver: context.driver,
+      direction: context.direction,
+      magnitude: context.magnitude,
+    });
+  }
+
+  #loadCache() {
+    try {
+      if (this.cachePath && existsSync(this.cachePath)) {
+        return new Map(
+          Object.entries(JSON.parse(readFileSync(this.cachePath, "utf-8"))),
+        );
+      }
+    } catch {
+      /* cache corrupt or missing */
+    }
+    return new Map();
+  }
+}
+
+/**
+ * Creates a ProseEngine with real dependencies wired.
+ * @param {object} options
+ * @param {string} options.cachePath - Path to .prose-cache.json
+ * @param {string} options.mode - "cached" | "generate" | "no-prose"
+ * @param {boolean} [options.strict] - Fail on cache miss
+ * @param {import('@forwardimpact/libllm').LlmApi} [options.llmApi] - LLM client
+ * @returns {ProseEngine}
+ */
+export function createProseEngine(options) {
+  const logger = createLogger("universe");
+  const promptLoader = new PromptLoader(join(__dirname, "..", "prompts"));
+  return new ProseEngine({ ...options, promptLoader, logger });
+}

package/index.js

@@ -0,0 +1,2 @@
+export { ProseEngine, createProseEngine } from "./engine/prose.js";
+export { PathwayGenerator, loadSchemas } from "./engine/pathway.js";

package/package.json

@@ -0,0 +1,32 @@
+{
+  "name": "@forwardimpact/libsyntheticprose",
+  "version": "0.1.1",
+  "description": "LLM-based prose and pathway generation for synthetic data",
+  "license": "Apache-2.0",
+  "repository": {
+    "type": "git",
+    "url": "https://github.com/forwardimpact/monorepo",
+    "directory": "libraries/libsyntheticprose"
+  },
+  "type": "module",
+  "main": "index.js",
+  "exports": {
+    ".": "./index.js",
+    "./prose": "./engine/prose.js",
+    "./pathway": "./engine/pathway.js"
+  },
+  "dependencies": {
+    "@forwardimpact/libutil": "^0.1.61",
+    "@forwardimpact/libtelemetry": "^0.1.23",
+    "@forwardimpact/libprompt": "^0.1.0",
+    "ajv": "^8.12.0",
+    "ajv-formats": "^2.1.1",
+    "yaml": "^2.3.0"
+  },
+  "engines": {
+    "node": ">=18.0.0"
+  },
+  "publishConfig": {
+    "access": "public"
+  }
+}

package/prompts/pathway/behaviour.js

@@ -0,0 +1,42 @@
+/**
+ * Prompt template for a single behaviour entity.
+ *
+ * @param {object} skeleton - Behaviour skeleton { id, name }
+ * @param {object} ctx - Universe context
+ * @param {object} schema - JSON schema for behaviour
+ * @returns {{ system: string, user: string }}
+ */
+export function buildBehaviourPrompt(skeleton, ctx, schema) {
+  return {
+    system: [
+      "You are an expert career framework author.",
+      "Output ONLY valid JSON. No markdown fences, no explanations.",
+      `The organization domain is: ${ctx.domain}.`,
+      `Industry: ${ctx.industry}.`,
+    ].join(" "),
+
+    user: [
+      "Generate a behaviour definition for a career framework.",
+      "",
+      "## JSON Schema (you MUST conform to this exactly)",
+      "```json",
+      JSON.stringify(schema, null, 2),
+      "```",
+      "",
+      `## Behaviour: "${skeleton.name}" (ID: ${skeleton.id})`,
+      "",
+      "## Instructions",
+      "- name: Use the provided name exactly.",
+      "- human.description: 2-3 sentences describing this behaviour.",
+      "- human.maturityDescriptions: One paragraph per maturity level",
+      "  (emerging, developing, practicing, role_modeling, exemplifying).",
+      '  Use second-person ("You..."). Each level must show clear',
+      "  progression in depth, consistency, and influence.",
+      "- agent.title: Short title (2-4 words) for how the agent applies this behaviour.",
+      "- agent.workingStyle: 1-2 sentences describing how the AI agent should embody",
+      "  this behaviour in its work style and communication.",
+      "",
+      "Output a single JSON object for this behaviour file.",
+    ].join("\n"),
+  };
+}

package/prompts/pathway/capability.js

@@ -0,0 +1,70 @@
+/**
+ * Prompt template for a single capability entity (with skills).
+ *
+ * @param {object} skeleton - Capability skeleton { id, name, skills, ordinalRank }
+ * @param {object} ctx - Universe context
+ * @param {object} schema - JSON schema for capability
+ * @returns {{ system: string, user: string }}
+ */
+export function buildCapabilityPrompt(skeleton, ctx, schema) {
+  return {
+    system: [
+      "You are an expert career framework author.",
+      "Output ONLY valid JSON. No markdown fences, no explanations.",
+      `The organization domain is: ${ctx.domain}.`,
+      `Industry: ${ctx.industry}.`,
+    ].join(" "),
+
+    user: [
+      "Generate a capability definition for a career framework.",
+      "",
+      "## JSON Schema (you MUST conform to this exactly)",
+      "```json",
+      JSON.stringify(schema, null, 2),
+      "```",
+      "",
+      `## Skeleton`,
+      `Capability ID: ${skeleton.id}`,
+      `Capability name: ${skeleton.name}`,
+      `Skills to define: ${skeleton.skills.join(", ")}`,
+      `Ordinal rank: ${skeleton.ordinalRank}`,
+      "",
+      "## Instructions",
+      "- id: Use the provided capability ID.",
+      "- name: Use the provided name.",
+      "- emojiIcon: A single emoji representing this capability.",
+      `- ordinalRank: ${skeleton.ordinalRank}`,
+      "- description: 1-2 sentences describing this capability area.",
+      "- professionalResponsibilities: One sentence per proficiency level",
+      "  (awareness through expert) describing IC expectations.",
+      "- managementResponsibilities: Same for management track.",
+      "- skills: For each skill ID listed above, generate:",
+      "  - id: Use the provided skill ID exactly.",
+      "  - name: Human-readable name (title case).",
+      "  - human.description: 2-3 sentences.",
+      "  - human.proficiencyDescriptions: One paragraph per level",
+      "    (awareness, foundational, working, practitioner, expert).",
+      '    Use second-person ("You..."). Each level must show clear',
+      "    progression in scope, autonomy, and complexity.",
+      "- For each skill, also generate an agent section:",
+      "  - agent.name: kebab-case name (e.g., 'code-review', 'data-modeling').",
+      "  - agent.description: 1 sentence describing what this agent skill provides.",
+      "  - agent.useWhen: 1 sentence describing when/why an agent should use this skill.",
+      "  - agent.stages: Object with ONLY the stages where this skill is meaningfully relevant.",
+      "    Not all skills need all 6 stages. Use these criteria:",
+      "      - specify: include if the skill informs what to build or constrains requirements",
+      "      - plan: include if the skill drives architecture or design decisions",
+      "      - onboard: include if the skill requires tooling, dependencies, or env setup",
+      "      - code: include if the skill is directly exercised during implementation",
+      "      - review: include if the skill has quality criteria to verify",
+      "      - deploy: include if the skill has production or operational concerns",
+      "    Each skill must have at least 2 stages. Omit stages where the skill has no specific guidance.",
+      "    Each stage has:",
+      "    - focus: 1 sentence — the primary focus for this skill in this stage.",
+      "    - readChecklist: Array of 2-3 items — steps to read/understand before acting.",
+      "    - confirmChecklist: Array of 2-3 items — items to verify after completing work.",
+      "",
+      "Output the JSON object for this single capability file.",
+    ].join("\n"),
+  };
+}

package/prompts/pathway/discipline.js

@@ -0,0 +1,61 @@
+/**
+ * Prompt template for a single discipline entity.
+ *
+ * @param {object} skeleton - Discipline skeleton from DSL
+ * @param {object} ctx - Universe context (includes skillIds, behaviourIds, trackIds)
+ * @param {object} schema - JSON schema for discipline
+ * @returns {{ system: string, user: string }}
+ */
+export function buildDisciplinePrompt(skeleton, ctx, schema) {
+  return {
+    system: [
+      "You are an expert career framework author.",
+      "Output ONLY valid JSON. No markdown fences, no explanations.",
+      `The organization domain is: ${ctx.domain}.`,
+      `Industry: ${ctx.industry}.`,
+    ].join(" "),
+
+    user: [
+      "Generate a discipline definition for a career framework.",
+      "",
+      "## JSON Schema (you MUST conform to this exactly)",
+      "```json",
+      JSON.stringify(schema, null, 2),
+      "```",
+      "",
+      `## Skeleton`,
+      `Discipline ID: ${skeleton.id}`,
+      `Role title: ${skeleton.roleTitle || skeleton.id.replace(/_/g, " ")}`,
+      `Specialization: ${skeleton.specialization || skeleton.roleTitle || skeleton.id.replace(/_/g, " ")}`,
+      `isProfessional: ${skeleton.isProfessional !== false}`,
+      `Core skills: ${(skeleton.core || []).join(", ")}`,
+      `Supporting skills: ${(skeleton.supporting || []).join(", ")}`,
+      `Broad skills: ${(skeleton.broad || []).join(", ")}`,
+      `Valid tracks: ${JSON.stringify(skeleton.validTracks || [null])}`,
+      "",
+      `## Available skill IDs: ${(ctx.skillIds || []).join(", ")}`,
+      `## Available behaviour IDs: ${(ctx.behaviourIds || []).join(", ")}`,
+      `## Available track IDs: ${(ctx.trackIds || []).join(", ")}`,
+      "",
+      "## Instructions",
+      "- specialization: Use the provided specialization or generate from role title.",
+      "- roleTitle: Use the provided role title.",
+      "- isProfessional: Use the provided value.",
+      "- isManagement: Set to true only for management disciplines.",
+      "- validTracks: Use the provided array. null means trackless/generalist is allowed.",
+      "- description: 2-3 sentences describing this discipline.",
+      "- coreSkills: Use the provided core skill IDs. Must all exist in available list.",
+      "- supportingSkills: Use the provided supporting skill IDs.",
+      "- broadSkills: Use the provided broad skill IDs.",
+      "- behaviourModifiers: Object mapping behaviour IDs to modifiers (-1, 0, or 1).",
+      "  Include 2-3 behaviour modifiers relevant to this discipline.",
+      "- human.roleSummary: 2-3 sentences describing this role. May use {roleTitle} or {specialization}.",
+      "- agent.identity: 1-2 sentences defining the AI coding agent's core identity.",
+      "  Frame as 'You are a {roleTitle} agent that...' May use {roleTitle} or {roleName} placeholder.",
+      "- agent.priority: 1 sentence stating the agent's top priority (e.g., code quality, system reliability).",
+      "- agent.constraints: 2-3 things the agent must avoid or never do.",
+      "",
+      "Output a single JSON object for this discipline.",
+    ].join("\n"),
+  };
+}

package/prompts/pathway/driver.js

@@ -0,0 +1,56 @@
+/**
+ * Prompt template for drivers.yaml — all drivers in a single call.
+ *
+ * @param {object[]} drivers - Driver skeletons from DSL
+ * @param {object} ctx - Universe context (includes skillIds, behaviourIds)
+ * @param {object} schema - JSON schema for drivers
+ * @returns {{ system: string, user: string }}
+ */
+export function buildDriverPrompt(drivers, ctx, schema) {
+  const driverList = drivers
+    .map((d) => {
+      const parts = [`  - id: ${d.id}, name: "${d.name}"`];
+      if (d.skills?.length) parts.push(`    skills: [${d.skills.join(", ")}]`);
+      if (d.behaviours?.length)
+        parts.push(`    behaviours: [${d.behaviours.join(", ")}]`);
+      return parts.join("\n");
+    })
+    .join("\n");
+
+  return {
+    system: [
+      "You are an expert career framework author.",
+      "Output ONLY valid JSON. No markdown fences, no explanations.",
+      `The organization domain is: ${ctx.domain}.`,
+      `Industry: ${ctx.industry}.`,
+    ].join(" "),
+
+    user: [
+      "Generate organizational driver definitions for a career framework.",
+      "",
+      "## JSON Schema (you MUST conform to this exactly)",
+      "```json",
+      JSON.stringify(schema, null, 2),
+      "```",
+      "",
+      "## Driver Skeletons",
+      driverList,
+      "",
+      `## Available skill IDs: ${(ctx.skillIds || []).join(", ")}`,
+      `## Available behaviour IDs: ${(ctx.behaviourIds || []).join(", ")}`,
+      "",
+      "## Instructions",
+      "- Output a JSON array of driver objects.",
+      "- For each driver:",
+      "  - id: Use the provided ID.",
+      "  - name: Use the provided name.",
+      "  - description: 2-3 sentences describing this organizational outcome.",
+      "  - contributingSkills: Use the skill IDs listed in the skeleton.",
+      "    All referenced skill IDs MUST be from the available list above.",
+      "  - contributingBehaviours: Use the behaviour IDs listed in the skeleton.",
+      "    All referenced behaviour IDs MUST be from the available list above.",
+      "",
+      "Output a JSON array.",
+    ].join("\n"),
+  };
+}

package/prompts/pathway/framework.js

@@ -0,0 +1,39 @@
+/**
+ * Prompt template for framework.yaml metadata.
+ *
+ * @param {object} skeleton - Framework skeleton from DSL
+ * @param {object} ctx - Universe context (domain, industry)
+ * @param {object} schema - JSON schema for framework entity
+ * @returns {{ system: string, user: string }}
+ */
+export function buildFrameworkPrompt(skeleton, ctx, schema) {
+  return {
+    system: [
+      "You are an expert career framework author.",
+      "Output ONLY valid JSON. No markdown fences, no explanations.",
+      `The organization domain is: ${ctx.domain}.`,
+      `Industry: ${ctx.industry}.`,
+    ].join(" "),
+
+    user: [
+      "Generate a framework metadata file for an engineering career framework.",
+      "",
+      "## JSON Schema (you MUST conform to this exactly)",
+      "```json",
+      JSON.stringify(schema, null, 2),
+      "```",
+      "",
+      "## Instructions",
+      '- title: A short, compelling title for this engineering pathway (e.g., "BioNova Engineering Pathway").',
+      "- emojiIcon: A single emoji representing engineering growth.",
+      '- tag: A short hashtag identifier (e.g., "#BioNova").',
+      "- description: 2-3 sentences describing the framework's purpose.",
+      `- distribution.siteUrl: Use "https://${ctx.domain}/pathway".`,
+      "- entityDefinitions: Provide definitions for these entity types:",
+      "  driver, skill, behaviour, discipline, level, track, job, agent, stage, tool.",
+      "  Each needs: title, emojiIcon, description (1 sentence).",
+      "",
+      "Output a single JSON object.",
+    ].join("\n"),
+  };
+}

package/prompts/pathway/level.js

@@ -0,0 +1,58 @@
+/**
+ * Prompt template for levels.yaml — all levels in a single call.
+ *
+ * @param {object[]} levels - Level skeletons from DSL
+ * @param {object} ctx - Universe context
+ * @param {object} schema - JSON schema for levels
+ * @returns {{ system: string, user: string }}
+ */
+export function buildLevelPrompt(levels, ctx, schema) {
+  const levelList = levels
+    .map(
+      (l) =>
+        `  - id: ${l.id}, professionalTitle: "${l.professionalTitle || ""}", rank: ${l.rank}, experience: "${l.experience || ""}"`,
+    )
+    .join("\n");
+
+  return {
+    system: [
+      "You are an expert career framework author.",
+      "Output ONLY valid JSON. No markdown fences, no explanations.",
+      `The organization domain is: ${ctx.domain}.`,
+      `Industry: ${ctx.industry}.`,
+    ].join(" "),
+
+    user: [
+      "Generate career level definitions for an engineering pathway.",
+      "",
+      "## JSON Schema (you MUST conform to this exactly)",
+      "```json",
+      JSON.stringify(schema, null, 2),
+      "```",
+      "",
+      "## Level Skeletons",
+      levelList,
+      "",
+      "## Instructions",
+      "- Output a JSON array of level objects.",
+      "- For each level, generate:",
+      "  - id: Use the provided ID (uppercase, e.g., J040).",
+      "  - professionalTitle: Use the provided title or generate one.",
+      "  - managementTitle: Generate a management-track equivalent.",
+      "  - ordinalRank: Use the provided rank.",
+      "  - typicalExperienceRange: Use the provided experience range.",
+      "  - qualificationSummary: 2-3 sentences describing qualifications.",
+      "    May use {typicalExperienceRange} placeholder.",
+      "  - baseSkillProficiencies: { primary, secondary, broad } using",
+      "    awareness/foundational/working/practitioner/expert.",
+      "    Increase across levels (L1→awareness, L5→expert for primary).",
+      "  - baseBehaviourMaturity: emerging/developing/practicing/role_modeling/exemplifying.",
+      "    Increase across levels.",
+      "  - expectations: { impactScope, autonomyExpectation, influenceScope, complexityHandled }.",
+      "    Each 1 sentence showing clear progression.",
+      "  - breadthCriteria: Only for rank >= 4. Object mapping proficiency → min count.",
+      "",
+      "Output a JSON array.",
+    ].join("\n"),
+  };
+}

package/prompts/pathway/stage.js

@@ -0,0 +1,45 @@
+/**
+ * Prompt template for stages.yaml — all stages in a single call.
+ *
+ * @param {string[]} stageIds - Stage ID list from DSL
+ * @param {object} ctx - Universe context
+ * @param {object} schema - JSON schema for stages
+ * @returns {{ system: string, user: string }}
+ */
+export function buildStagePrompt(stageIds, ctx, schema) {
+  return {
+    system: [
+      "You are an expert career framework author.",
+      "Output ONLY valid JSON. No markdown fences, no explanations.",
+      `The organization domain is: ${ctx.domain}.`,
+      `Industry: ${ctx.industry}.`,
+    ].join(" "),
+
+    user: [
+      "Generate engineering lifecycle stage definitions.",
+      "",
+      "## JSON Schema (you MUST conform to this exactly)",
+      "```json",
+      JSON.stringify(schema, null, 2),
+      "```",
+      "",
+      `## Stage IDs: ${stageIds.join(", ")}`,
+      "",
+      "## Instructions",
+      "- Output a JSON array of stage objects.",
+      "- For each stage ID, generate:",
+      "  - id: The stage ID (must be one of: specify, plan, onboard, code, review, deploy).",
+      '  - name: Human-readable name (e.g., "Specify", "Plan").',
+      "  - emojiIcon: A single emoji for this stage.",
+      '  - description: 2-3 sentences in second person ("You...").',
+      "  - summary: 1 sentence in third person.",
+      "  - handoffs: Array of transitions to other stages, each with:",
+      "    targetStage, label (button text), prompt (instructions for next stage).",
+      "  - constraints: 2-3 restrictions on behaviour in this stage.",
+      "  - readChecklist: 2-4 Read-Then-Do steps.",
+      "  - confirmChecklist: 2-4 Do-Then-Confirm items.",
+      "",
+      "Output a JSON array.",
+    ].join("\n"),
+  };
+}

package/prompts/pathway/track.js

@@ -0,0 +1,51 @@
+/**
+ * Prompt template for a single track entity.
+ *
+ * @param {object} skeleton - Track skeleton { id, name }
+ * @param {object} ctx - Universe context (includes capabilityIds, behaviourIds)
+ * @param {object} schema - JSON schema for track
+ * @returns {{ system: string, user: string }}
+ */
+export function buildTrackPrompt(skeleton, ctx, schema) {
+  return {
+    system: [
+      "You are an expert career framework author.",
+      "Output ONLY valid JSON. No markdown fences, no explanations.",
+      `The organization domain is: ${ctx.domain}.`,
+      `Industry: ${ctx.industry}.`,
+    ].join(" "),
+
+    user: [
+      "Generate a track definition for a career framework.",
+      "",
+      "## JSON Schema (you MUST conform to this exactly)",
+      "```json",
+      JSON.stringify(schema, null, 2),
+      "```",
+      "",
+      `## Skeleton`,
+      `Track ID: ${skeleton.id}`,
+      `Track name: ${skeleton.name}`,
+      "",
+      `## Available capability IDs: ${(ctx.capabilityIds || []).join(", ")}`,
+      `## Available behaviour IDs: ${(ctx.behaviourIds || []).join(", ")}`,
+      "",
+      "## Instructions",
+      "- name: Use the provided name exactly.",
+      "- description: 2-3 sentences describing this track's focus.",
+      "- roleContext: 1-2 sentences contextualizing the role for job listings.",
+      "- skillModifiers: Object mapping capability IDs to integer modifiers.",
+      "  Use values from -1 to 1. Include modifiers for capabilities most",
+      "  affected by this track specialization.",
+      "- behaviourModifiers: Object mapping behaviour IDs to integer modifiers.",
+      "  Include 1-2 relevant modifiers.",
+      "- assessmentWeights: { skillWeight, behaviourWeight } summing to 1.",
+      "- agent.identity: 1 sentence identity override for the agent when working in this track.",
+      "  May use {roleTitle} placeholder. Example: 'You specialize in platform infrastructure.'",
+      "- agent.priority: 1 sentence stating the track-specific priority.",
+      "- agent.constraints: 1-2 additional constraints specific to this track.",
+      "",
+      "Output a single JSON object for this track.",
+    ].join("\n"),
+  };
+}

package/prompts/prose-system.prompt.md

@@ -0,0 +1,2 @@
+You are a technical writer for a pharmaceutical company. Generate concise,
+realistic content. Output the text only, no explanations or markdown formatting.

package/prompts/prose-user.prompt.md

@@ -0,0 +1,6 @@
+Write {{length}} of {{tone}} prose about: {{topic}}. {{#domain}} Company domain:
+{{domain}}. {{/domain}} {{#role}} Written from the perspective of: {{role}}.
+{{/role}} {{#audience}} Target audience: {{audience}}. {{/audience}}
+{{#scenario}} Context: during "{{scenario}}", the {{driver}} driver is
+{{direction}} (magnitude: {{magnitude}}). {{/scenario}} Output the text only, no
+explanations.