feat: phase 1 of AI pipeline hardening — single LLM client + tier-aware models

Implements phase 1 of AI_PIPELINE_HARDENING_PLAN.md. Every Anthropic call
now goes through one module that owns retry, timeout, abort, structured-
output parsing, schema validation, and best-effort call telemetry.

* src/lib/llm.js — single callLLM entry point. Resolves model per tier
  (fast / standard / reasoning) with admin:model legacy fallback for the
  standard tier; 60s default timeout via AbortController; balanced-brace
  JSON extraction; LLMHttpError, LLMTruncatedError, LLMOutputError, and
  LLMValidationError surface clearly distinct failure modes.
* src/lib/llmRetry.js — exponential backoff with full jitter, retries
  only on transient HTTP statuses, honours Retry-After up to 60s, never
  retries on AbortError.
* src/lib/llmSchemas.js — Zod schemas for every structured task plus
  normalizeHandbookResult (collapses legacy "executes" relations into
  the canonical "executed_by" vocabulary).
* src/lib/api.js — thin shim over callLLM so existing callers (extraction
  pipeline, learning, quiz, R42, knowledge graph) keep working unchanged.
* src/lib/__tests__/ — 32 Vitest cases covering parse paths, error
  surfaces, simulation mode, model resolution, and schema validation.
* src/pages/Admin/index.jsx — three model inputs (fast / standard /
  reasoning) replacing the single legacy field; legacy value falls back
  for the standard tier so existing overrides survive.

Adds Zod and Vitest, plus an "npm run test" script.

Also cleans up the pre-existing repo-wide ESLint failures so phase 1's
"npm run lint passes" acceptance criterion can be checked: drops unused
React imports across the JSX tree (React 19 JSX runtime auto-imports),
attaches cause to rethrown errors in the service modules, ignores
pb_migrations in the ESLint config (PocketBase JSVM globals), and
removes one dead handleCreateCustom function in Leren.jsx. A real
behaviour bug surfaced in Testen.jsx — the quiz timer captured a stale
finishQuiz via setInterval closure; now updated via finishQuizRef so the
timer always invokes the latest callback.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>

src/lib/api.js

@@ -1,135 +1,39 @@
-import { storage } from './storage';
-
 /**
- * Anthropic API Service
- * Handles communication with the /v1/messages endpoint via Nginx proxy.
+ * Back-compatibility shim for the legacy `anthropicApi` interface.
+ *
+ * All real work lives in `./llm.js`. Existing callers (extractionPipeline,
+ * learningService, testService, KnowledgeGraph, useChat) keep working
+ * unchanged; new code should import `callLLM` from `./llm.js` directly.
  */
 
-const DEFAULT_MODEL = 'claude-sonnet-4-20250514';
+import { callLLM } from './llm';
 
 export const anthropicApi = {
-  async generateContent(systemPrompt, userMessage, maxRetries = 1) {
-    // Check if simulation mode is on
-    const useSimulation = storage.get('admin:use_simulation') === true;
-    if (useSimulation) {
-      console.log('[API] Simulation mode active. Mock data will be returned.');
-      return await simulateResponse();
-    }
+  async generateContent(systemPrompt, userMessage /*, maxRetries */) {
+    const { text } = await callLLM({
+      task: 'legacy.generateContent',
+      tier: 'standard',
+      system: systemPrompt,
+      user: userMessage,
+      maxTokens: 8192,
+      temperature: 0,
+    });
+    return text;
+  },
 
-    // The API key is now securely injected by the Caddy reverse proxy via environment variables.
-
-    // Model is configurable from Admin > Settings, defaults to the original spec model
-    const model = storage.get('admin:model') || DEFAULT_MODEL;
-    console.log(`[API] Calling with model: ${model}`);
-
-    let retries = 0;
-    while (retries <= maxRetries) {
-      try {
-        const response = await fetch('/api/anthropic/v1/messages', {
-          method: 'POST',
-          headers: {
-            'Content-Type': 'application/json',
-            'anthropic-version': '2023-06-01',
-          },
-          body: JSON.stringify({
-            model: model,
-            max_tokens: 8192,
-            temperature: 0,
-            system: systemPrompt,
-            messages: [{ role: 'user', content: userMessage }]
-          })
-        });
-
-        if (!response.ok) {
-          const errData = await response.json().catch(() => ({}));
-          console.error('[API] Error response:', errData);
-          throw new Error(`API Error: ${response.status} ${response.statusText} - ${JSON.stringify(errData)}`);
-        }
-
-        // Detect auth portal session expiry: the portal returns HTML instead of JSON
-        const contentType = response.headers.get('content-type') || '';
-        if (!contentType.includes('application/json')) {
-          throw new Error('Your session has expired. Please refresh the page and log in again.');
-        }
-
-        const data = await response.json();
-        return data.content[0].text;
-      } catch (error) {
-        console.error('API call failed:', error);
-        retries++;
-        if (retries > maxRetries) throw error;
-        await new Promise(r => setTimeout(r, 1000));
-      }
-    }
-  }
+  async chat(systemPrompt, messages, opts = {}) {
+    const r = await callLLM({
+      task: 'legacy.chat',
+      tier: 'standard',
+      system: systemPrompt,
+      messages,
+      tools: opts.tools,
+      maxTokens: 1024,
+      temperature: 0.3,
+    });
+    const content = [];
+    if (r.text) content.push({ type: 'text', text: r.text });
+    for (const tu of r.toolUses) content.push({ type: 'tool_use', name: tu.name, input: tu.input });
+    return { content, stop_reason: r.stopReason };
+  },
 };
-
-/**
- * Multi-turn chat with optional tool use.
- * Returns the raw Anthropic response so callers can read both `text` and
- * `tool_use` content blocks.
- *
- * @param {string} systemPrompt
- * @param {Array<{role: 'user'|'assistant', content: string}>} messages
- * @param {{tools?: Array}} opts
- * @returns {Promise<{content: Array, stop_reason: string}>}
- */
-anthropicApi.chat = async function chat(systemPrompt, messages, opts = {}) {
-  const useSimulation = storage.get('admin:use_simulation') === true;
-  if (useSimulation) {
-    await new Promise(r => setTimeout(r, 600));
-    return {
-      content: [{
-        type: 'text',
-        text: 'Simulatiemodus staat aan — vraag een beheerder om Simulation Mode uit te zetten in Admin → Settings om met R42 te chatten.',
-      }],
-      stop_reason: 'end_turn',
-    };
-  }
-
-  const model = storage.get('admin:model') || DEFAULT_MODEL;
-
-  const body = {
-    model,
-    max_tokens: 1024,
-    system: systemPrompt,
-    messages,
-  };
-  if (opts.tools && opts.tools.length) body.tools = opts.tools;
-
-  const response = await fetch('/api/anthropic/v1/messages', {
-    method: 'POST',
-    headers: {
-      'Content-Type': 'application/json',
-      'anthropic-version': '2023-06-01',
-    },
-    body: JSON.stringify(body),
-  });
-
-  if (!response.ok) {
-    const errData = await response.json().catch(() => ({}));
-    throw new Error(`API Error: ${response.status} ${response.statusText} - ${JSON.stringify(errData)}`);
-  }
-
-  const contentType = response.headers.get('content-type') || '';
-  if (!contentType.includes('application/json')) {
-    throw new Error('Your session has expired. Please refresh the page and log in again.');
-  }
-
-  return await response.json();
-};
-
-async function simulateResponse() {
-  await new Promise(r => setTimeout(r, 2000));
-  return JSON.stringify({
-    topics: [
-      { id: "radicale-transparantie", label: "Radicale Transparantie", type: "concept", description: "De kernwaarde van Respellion waarbij alle informatie publiek toegankelijk is." },
-      { id: "kennisbeheer", label: "Kennisbeheer", type: "process", description: "Het proces van het vastleggen en ontsluiten van organisatiekennis." },
-      { id: "wekelijkse-sessie", label: "Wekelijkse Leersessie", type: "process", description: "Elke week leren medewerkers via AI-gegenereerde vragen en quizzen." }
-    ],
-    relations: [
-      { source: "kennisbeheer", target: "radicale-transparantie", type: "depends_on" },
-      { source: "wekelijkse-sessie", target: "kennisbeheer", type: "part_of" }
-    ]
-  });
-}