fix(ai): wrap provider calls in langchain and restore arithmetic direct replies

Tasks.md

@@ -17,6 +17,7 @@ Last updated: 2026-02-24
 | T-009 | Open source eval framework contribution | In Review | `@ghostfolio/finance-agent-evals` package scaffold + dataset export + smoke/pack checks | openai/evals PR #1625 + langchain PR #35421 |
 | T-010 | Chat history persistence and simple direct-query handling | Complete | `apps/client/src/app/pages/portfolio/analysis/ai-chat-panel/ai-chat-panel.component.spec.ts`, `apps/api/src/app/endpoints/ai/ai-agent.utils.spec.ts`, `apps/api/src/app/endpoints/ai/ai.service.spec.ts` | Local implementation |
 | T-011 | Per-LLM LangSmith invocation tracing + production tracing env enablement | Complete | `apps/api/src/app/endpoints/ai/ai-observability.service.spec.ts`, `apps/api/src/app/endpoints/ai/ai.service.spec.ts`, `apps/api/src/app/endpoints/ai/ai-performance.spec.ts`, `apps/api/src/app/endpoints/ai/evals/mvp-eval.runner.spec.ts`, `apps/api/src/app/endpoints/ai/evals/ai-quality-eval.spec.ts` | Local implementation + Railway variable update |
+| T-012 | LangChain wrapper enforcement for provider calls + arithmetic direct-response correction | Complete | `apps/api/src/app/endpoints/ai/ai.service.spec.ts`, `apps/api/src/app/endpoints/ai/ai-agent.utils.spec.ts`, `npm run test:ai` | Local implementation |
 
 ## Notes
 
@@ -37,3 +38,4 @@ Last updated: 2026-02-24
 - Tool gating hardening (2026-02-24): planner unknown-intent fallback changed to no-tools, executor policy gate added (`direct|tools|clarify`), and policy metrics emitted via verification and observability logs.
 - Chat persistence + simple direct-query handling (2026-02-24): client chat panel now restores/persists session + bounded message history via localStorage and policy no-tool prompts now return assistant capability guidance for queries like "Who are you?".
 - Per-LLM LangSmith invocation tracing (2026-02-24): each provider call now records an explicit LangSmith `llm` run (provider/model/query/session/response metadata), and production Railway env now has tracing variables enabled.
+- Direct arithmetic no-tool behavior fix (2026-02-24): simple arithmetic prompts now return computed answers (for example `2+2 = 4`) instead of generic capability guidance.

apps/api/src/app/endpoints/ai/ai-agent.policy.utils.ts

@@ -28,6 +28,7 @@ const GREETING_ONLY_PATTERN =
 const SIMPLE_ARITHMETIC_QUERY_PATTERN =
   /^\s*(?:what(?:'s| is)\s+)?[-+*/().\d\s%=]+\??\s*$/i;
 const SIMPLE_ARITHMETIC_OPERATOR_PATTERN = /[+\-*/]/;
+const SIMPLE_ARITHMETIC_PREFIX_PATTERN = /^\s*(?:what(?:'s| is)\s+)?/i;
 const SIMPLE_ASSISTANT_QUERY_PATTERNS = [
   /^\s*(?:who are you|what are you|what can you do)\s*[!.?]*\s*$/i,
   /^\s*(?:how do you work|how (?:can|do) i use (?:you|this))\s*[!.?]*\s*$/i,
@@ -94,6 +95,203 @@ function isNoToolDirectQuery(query: string) {
   );
 }
 
+function formatNumericResult(value: number) {
+  if (Math.abs(value) < Number.EPSILON) {
+    return '0';
+  }
+
+  if (Number.isInteger(value)) {
+    return value.toString();
+  }
+
+  return value.toFixed(6).replace(/\.?0+$/, '');
+}
+
+function evaluateArithmeticExpression(expression: string) {
+  let cursor = 0;
+
+  const skipWhitespace = () => {
+    while (cursor < expression.length && /\s/.test(expression[cursor])) {
+      cursor += 1;
+    }
+  };
+
+  const parseNumber = () => {
+    skipWhitespace();
+    const start = cursor;
+    let dotCount = 0;
+
+    while (cursor < expression.length) {
+      const token = expression[cursor];
+
+      if (token >= '0' && token <= '9') {
+        cursor += 1;
+        continue;
+      }
+
+      if (token === '.') {
+        dotCount += 1;
+
+        if (dotCount > 1) {
+          return undefined;
+        }
+
+        cursor += 1;
+        continue;
+      }
+
+      break;
+    }
+
+    if (start === cursor) {
+      return undefined;
+    }
+
+    const parsed = Number(expression.slice(start, cursor));
+
+    return Number.isFinite(parsed) ? parsed : undefined;
+  };
+
+  const parseFactor = (): number | undefined => {
+    skipWhitespace();
+
+    if (expression[cursor] === '+') {
+      cursor += 1;
+      return parseFactor();
+    }
+
+    if (expression[cursor] === '-') {
+      cursor += 1;
+      const nested = parseFactor();
+
+      return nested === undefined ? undefined : -nested;
+    }
+
+    if (expression[cursor] === '(') {
+      cursor += 1;
+      const nested = parseExpression();
+
+      skipWhitespace();
+      if (nested === undefined || expression[cursor] !== ')') {
+        return undefined;
+      }
+
+      cursor += 1;
+      return nested;
+    }
+
+    return parseNumber();
+  };
+
+  const parseTerm = (): number | undefined => {
+    let value = parseFactor();
+
+    if (value === undefined) {
+      return undefined;
+    }
+
+    while (true) {
+      skipWhitespace();
+      const operator = expression[cursor];
+
+      if (operator !== '*' && operator !== '/') {
+        break;
+      }
+
+      cursor += 1;
+      const right = parseFactor();
+
+      if (right === undefined) {
+        return undefined;
+      }
+
+      if (operator === '*') {
+        value *= right;
+      } else {
+        if (Math.abs(right) < Number.EPSILON) {
+          return undefined;
+        }
+
+        value /= right;
+      }
+    }
+
+    return value;
+  };
+
+  const parseExpression = (): number | undefined => {
+    let value = parseTerm();
+
+    if (value === undefined) {
+      return undefined;
+    }
+
+    while (true) {
+      skipWhitespace();
+      const operator = expression[cursor];
+
+      if (operator !== '+' && operator !== '-') {
+        break;
+      }
+
+      cursor += 1;
+      const right = parseTerm();
+
+      if (right === undefined) {
+        return undefined;
+      }
+
+      if (operator === '+') {
+        value += right;
+      } else {
+        value -= right;
+      }
+    }
+
+    return value;
+  };
+
+  const result = parseExpression();
+
+  skipWhitespace();
+
+  if (result === undefined || cursor !== expression.length || !Number.isFinite(result)) {
+    return undefined;
+  }
+
+  return result;
+}
+
+function evaluateSimpleArithmetic(query: string) {
+  const normalized = query.trim();
+
+  if (
+    !SIMPLE_ARITHMETIC_QUERY_PATTERN.test(normalized) ||
+    !SIMPLE_ARITHMETIC_OPERATOR_PATTERN.test(normalized) ||
+    !/\d/.test(normalized)
+  ) {
+    return undefined;
+  }
+
+  const expression = normalized
+    .replace(SIMPLE_ARITHMETIC_PREFIX_PATTERN, '')
+    .replace(/\?+$/, '')
+    .replace(/=/g, '')
+    .trim();
+
+  if (!expression) {
+    return undefined;
+  }
+
+  const result = evaluateArithmeticExpression(expression);
+
+  if (result === undefined) {
+    return undefined;
+  }
+
+  return `${expression} = ${formatNumericResult(result)}`;
+}
+
 export function applyToolExecutionPolicy({
   plannedTools,
   query
@@ -188,9 +386,11 @@ export function applyToolExecutionPolicy({
 }
 
 export function createPolicyRouteResponse({
-  policyDecision
+  policyDecision,
+  query
 }: {
   policyDecision: AiAgentToolPolicyDecision;
+  query?: string;
 }) {
   if (policyDecision.route === 'clarify') {
     if (policyDecision.blockReason === 'needs_confirmation') {
@@ -204,6 +404,12 @@ export function createPolicyRouteResponse({
     policyDecision.route === 'direct' &&
     policyDecision.blockReason === 'no_tool_query'
   ) {
+    const arithmeticResult = query ? evaluateSimpleArithmetic(query) : undefined;
+
+    if (arithmeticResult) {
+      return arithmeticResult;
+    }
+
     return `I am your Ghostfolio AI assistant. I can help with portfolio analysis, concentration risk, market prices, rebalancing ideas, and stress scenarios. Try: "Show my top holdings" or "What is my concentration risk?".`;
   }
 

apps/api/src/app/endpoints/ai/ai-agent.utils.spec.ts

@@ -83,6 +83,19 @@ describe('AiAgentUtils', () => {
     );
   });
 
+  it('returns deterministic arithmetic result for direct no-tool arithmetic query', () => {
+    const decision = applyToolExecutionPolicy({
+      plannedTools: [],
+      query: '2+2'
+    });
+
+    expect(decision.route).toBe('direct');
+    expect(decision.toolsToExecute).toEqual([]);
+    expect(createPolicyRouteResponse({ policyDecision: decision, query: '2+2' })).toBe(
+      '2+2 = 4'
+    );
+  });
+
   it('keeps finance-intent prompts on clarify route even with capability phrasing', () => {
     const decision = applyToolExecutionPolicy({
       plannedTools: [],

apps/api/src/app/endpoints/ai/ai.service.spec.ts

@@ -274,6 +274,23 @@ describe('AiService', () => {
     );
   });
 
+  it('returns arithmetic response on direct no-tool arithmetic query', async () => {
+    redisCacheService.get.mockResolvedValue(undefined);
+    const generateTextSpy = jest.spyOn(subject, 'generateText');
+
+    const result = await subject.chat({
+      languageCode: 'en',
+      query: '2+2',
+      sessionId: 'session-arithmetic-route',
+      userCurrency: 'USD',
+      userId: 'user-arithmetic-route'
+    });
+
+    expect(result.answer).toBe('2+2 = 4');
+    expect(result.toolCalls).toEqual([]);
+    expect(generateTextSpy).not.toHaveBeenCalled();
+  });
+
   it('runs rebalance and stress test tools for portfolio scenario prompts', async () => {
     portfolioService.getDetails.mockResolvedValue({
       holdings: {

apps/api/src/app/endpoints/ai/ai.service.ts

@@ -10,6 +10,7 @@ import { Filter } from '@ghostfolio/common/interfaces';
 import type { AiPromptMode } from '@ghostfolio/common/types';
 import { Injectable } from '@nestjs/common';
 import { createOpenRouter } from '@openrouter/ai-sdk-provider';
+import { RunnableLambda } from '@langchain/core/runnables';
 import { generateText } from 'ai';
 import { randomUUID } from 'node:crypto';
 import {
@@ -85,31 +86,71 @@ export class AiService {
       provider: string;
       run: () => Promise<{ text?: string }>;
     }) => {
-      const startedAt = Date.now();
-      let invocationError: unknown;
-      let responseText: string | undefined;
-
-      try {
-        const response = await run();
-        responseText = response?.text;
+      const invocationRunnable = RunnableLambda.from(
+        async ({
+          model: runnableModel,
+          prompt: runnablePrompt,
+          provider: runnableProvider,
+          query,
+          sessionId,
+          userId
+        }: {
+          model: string;
+          prompt: string;
+          provider: string;
+          query?: string;
+          sessionId?: string;
+          userId?: string;
+        }) => {
+          const startedAt = Date.now();
+          let invocationError: unknown;
+          let responseText: string | undefined;
+
+          try {
+            const response = await run();
+            responseText = response?.text;
+
+            return response;
+          } catch (error) {
+            invocationError = error;
+            throw error;
+          } finally {
+            void this.aiObservabilityService.recordLlmInvocation({
+              durationInMs: Date.now() - startedAt,
+              error: invocationError,
+              model: runnableModel,
+              prompt: runnablePrompt,
+              provider: runnableProvider,
+              query,
+              responseText,
+              sessionId,
+              userId
+            });
+          }
+        }
+      );
 
-        return response;
-      } catch (error) {
-        invocationError = error;
-        throw error;
-      } finally {
-        void this.aiObservabilityService.recordLlmInvocation({
-          durationInMs: Date.now() - startedAt,
-          error: invocationError,
+      return invocationRunnable.invoke(
+        {
           model,
           prompt,
           provider,
           query: traceContext?.query,
-          responseText,
           sessionId: traceContext?.sessionId,
           userId: traceContext?.userId
-        });
-      }
+        },
+        {
+          metadata: {
+            model,
+            provider,
+            query: traceContext?.query ?? '',
+            sessionId: traceContext?.sessionId ?? '',
+            userId: traceContext?.userId ?? ''
+          },
+          runName: `ghostfolio_ai_llm_${provider}`,
+          tags: ['ghostfolio-ai', 'llm-invocation', provider]
+        }
+      );
     };
 
     if (zAiGlmApiKey) {
@@ -393,7 +434,8 @@ export class AiService {
       });
 
       let answer = createPolicyRouteResponse({
-        policyDecision
+        policyDecision,
+        query: normalizedQuery
       });
 
       if (policyDecision.route === 'tools') {

package-lock.json

@@ -26,6 +26,7 @@
         "@internationalized/number": "3.6.5",
         "@ionic/angular": "8.7.8",
         "@keyv/redis": "4.4.0",
+        "@langchain/core": "^0.3.80",
         "@nestjs/bull": "11.0.4",
         "@nestjs/cache-manager": "3.0.1",
         "@nestjs/common": "11.1.8",
@@ -3633,6 +3634,12 @@
       "devOptional": true,
       "license": "(Apache-2.0 AND BSD-3-Clause)"
     },
+    "node_modules/@cfworker/json-schema": {
+      "version": "4.1.1",
+      "resolved": "https://registry.npmjs.org/@cfworker/json-schema/-/json-schema-4.1.1.tgz",
+      "integrity": "sha512-gAmrUZSGtKc3AiBL71iNWxDsyUC5uMaKKGdvzYsBoTW/xi42JQHl7eKV2OYzCUqvc+D2RCcf7EXY2iCyFIk6og==",
+      "license": "MIT"
+    },
     "node_modules/@chevrotain/cst-dts-gen": {
       "version": "11.0.3",
       "resolved": "https://registry.npmjs.org/@chevrotain/cst-dts-gen/-/cst-dts-gen-11.0.3.tgz",
@@ -6292,6 +6299,150 @@
       "integrity": "sha512-M5UknZPHRu3DEDWoipU6sE8PdkZ6Z/S+v4dD+Ke8IaNlpdSQah50lz1KtcFBa2vsdOnwbbnxJwVM4wty6udA5w==",
       "license": "MIT"
     },
+    "node_modules/@langchain/core": {
+      "version": "0.3.80",
+      "resolved": "https://registry.npmjs.org/@langchain/core/-/core-0.3.80.tgz",
+      "integrity": "sha512-vcJDV2vk1AlCwSh3aBm/urQ1ZrlXFFBocv11bz/NBUfLWD5/UDNMzwPdaAd2dKvNmTWa9FM2lirLU3+JCf4cRA==",
+      "license": "MIT",
+      "dependencies": {
+        "@cfworker/json-schema": "^4.0.2",
+        "ansi-styles": "^5.0.0",
+        "camelcase": "6",
+        "decamelize": "1.2.0",
+        "js-tiktoken": "^1.0.12",
+        "langsmith": "^0.3.67",
+        "mustache": "^4.2.0",
+        "p-queue": "^6.6.2",
+        "p-retry": "4",
+        "uuid": "^10.0.0",
+        "zod": "^3.25.32",
+        "zod-to-json-schema": "^3.22.3"
+      },
+      "engines": {
+        "node": ">=18"
+      }
+    },
+    "node_modules/@langchain/core/node_modules/@types/retry": {
+      "version": "0.12.0",
+      "resolved": "https://registry.npmjs.org/@types/retry/-/retry-0.12.0.tgz",
+      "integrity": "sha512-wWKOClTTiizcZhXnPY4wikVAwmdYHp8q6DmC+EJUzAMsycb7HB32Kh9RN4+0gExjmPmZSAQjgURXIGATPegAvA==",
+      "license": "MIT"
+    },
+    "node_modules/@langchain/core/node_modules/ansi-styles": {
+      "version": "5.2.0",
+      "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-5.2.0.tgz",
+      "integrity": "sha512-Cxwpt2SfTzTtXcfOlzGEee8O+c+MmUgGrNiBcXnuWxuFJHe6a5Hz7qwhwe5OgaSYI0IJvkLqWX1ASG+cJOkEiA==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=10"
+      },
+      "funding": {
+        "url": "https://github.com/chalk/ansi-styles?sponsor=1"
+      }
+    },
+    "node_modules/@langchain/core/node_modules/camelcase": {
+      "version": "6.3.0",
+      "resolved": "https://registry.npmjs.org/camelcase/-/camelcase-6.3.0.tgz",
+      "integrity": "sha512-Gmy6FhYlCY7uOElZUSbxo2UCDH8owEk996gkbrpsgGtrJLM3J7jGxl9Ic7Qwwj4ivOE5AWZWRMecDdF7hqGjFA==",
+      "license": "MIT",
+      "engines": {
+        "node": ">=10"
+      },
+      "funding": {
+        "url": "https://github.com/sponsors/sindresorhus"
+      }
+    },
+    "node_modules/@langchain/core/node_modules/chalk": {
+      "version": "4.1.2",
+      "resolved": "https://registry.npmjs.org/chalk/-/chalk-4.1.2.tgz",
+      "integrity": "sha512-oKnbhFyRIXpUuez8iBMmyEa4nbj4IOQyuhc/wy9kY7/WVPcwIO9VA668Pu8RkO7+0G76SLROeyw9CpQ061i4mA==",
+      "license": "MIT",
+      "dependencies": {
+        "ansi-styles": "^4.1.0",
+        "supports-color": "^7.1.0"
+      },
+      "engines": {
+        "node": ">=10"
+      },
+      "funding": {
+        "url": "https://github.com/chalk/chalk?sponsor=1"
+      }
+    },
+    "node_modules/@langchain/core/node_modules/chalk/node_modules/ansi-styles": {
+      "version": "4.3.0",
+      "resolved": "https://registry.npmjs.org/ansi-styles/-/ansi-styles-4.3.0.tgz",
+      "integrity": "sha512-zbB9rCJAT1rbjiVDb2hqKFHNYLxgtk8NURxZ3IZwD3F6NtxbXZQCnnSi1Lkx+IDohdPlFp222wVALIheZJQSEg==",
+      "license": "MIT",
+      "dependencies": {
+        "color-convert": "^2.0.1"
+      },
+      "engines": {
+        "node": ">=8"
+      },
+      "funding": {
+        "url": "https://github.com/chalk/ansi-styles?sponsor=1"
+      }
+    },
+    "node_modules/@langchain/core/node_modules/langsmith": {
+      "version": "0.3.87",
+      "resolved": "https://registry.npmjs.org/langsmith/-/langsmith-0.3.87.tgz",
+      "integrity": "sha512-XXR1+9INH8YX96FKWc5tie0QixWz6tOqAsAKfcJyPkE0xPep+NDz0IQLR32q4bn10QK3LqD2HN6T3n6z1YLW7Q==",
+      "license": "MIT",
+      "dependencies": {
+        "@types/uuid": "^10.0.0",
+        "chalk": "^4.1.2",
+        "console-table-printer": "^2.12.1",
+        "p-queue": "^6.6.2",
+        "semver": "^7.6.3",
+        "uuid": "^10.0.0"
+      },
+      "peerDependencies": {
+        "@opentelemetry/api": "*",
+        "@opentelemetry/exporter-trace-otlp-proto": "*",
+        "@opentelemetry/sdk-trace-base": "*",
+        "openai": "*"
+      },
+      "peerDependenciesMeta": {
+        "@opentelemetry/api": {
+          "optional": true
+        },
+        "@opentelemetry/exporter-trace-otlp-proto": {
+          "optional": true
+        },
+        "@opentelemetry/sdk-trace-base": {
+          "optional": true
+        },
+        "openai": {
+          "optional": true
+        }
+      }
+    },
+    "node_modules/@langchain/core/node_modules/p-retry": {
+      "version": "4.6.2",
+      "resolved": "https://registry.npmjs.org/p-retry/-/p-retry-4.6.2.tgz",
+      "integrity": "sha512-312Id396EbJdvRONlngUx0NydfrIQ5lsYu0znKVUzVvArzEIt08V1qhtyESbGVd1FGX7UKtiFp5uwKZdM8wIuQ==",
+      "license": "MIT",
+      "dependencies": {
+        "@types/retry": "0.12.0",
+        "retry": "^0.13.1"
+      },
+      "engines": {
+        "node": ">=8"
+      }
+    },
+    "node_modules/@langchain/core/node_modules/uuid": {
+      "version": "10.0.0",
+      "resolved": "https://registry.npmjs.org/uuid/-/uuid-10.0.0.tgz",
+      "integrity": "sha512-8XkAphELsDnEGrDxUOHB3RGvXz6TeuYSGEZBOjtTtPm2lwhGBjLgOzLHB63IUWfBpNucQjND6d3AOudO+H3RWQ==",
+      "funding": [
+        "https://github.com/sponsors/broofa",
+        "https://github.com/sponsors/ctavan"
+      ],
+      "license": "MIT",
+      "bin": {
+        "uuid": "dist/bin/uuid"
+      }
+    },
     "node_modules/@leichtgewicht/ip-codec": {
       "version": "2.0.5",
       "resolved": "https://registry.npmjs.org/@leichtgewicht/ip-codec/-/ip-codec-2.0.5.tgz",
@@ -24650,6 +24801,15 @@
       "license": "MIT",
       "peer": true
     },
+    "node_modules/js-tiktoken": {
+      "version": "1.0.21",
+      "resolved": "https://registry.npmjs.org/js-tiktoken/-/js-tiktoken-1.0.21.tgz",
+      "integrity": "sha512-biOj/6M5qdgx5TKjDnFT1ymSpM5tbd3ylwDtrQvFQSu0Z7bBYko2dF+W/aUkXUPuk6IVpRxk/3Q2sHOzGlS36g==",
+      "license": "MIT",
+      "dependencies": {
+        "base64-js": "^1.5.1"
+      }
+    },
     "node_modules/js-tokens": {
       "version": "4.0.0",
       "resolved": "https://registry.npmjs.org/js-tokens/-/js-tokens-4.0.0.tgz",
@@ -26735,6 +26895,15 @@
         "multicast-dns": "cli.js"
       }
     },
+    "node_modules/mustache": {
+      "version": "4.2.0",
+      "resolved": "https://registry.npmjs.org/mustache/-/mustache-4.2.0.tgz",
+      "integrity": "sha512-71ippSywq5Yb7/tVYyGbkBggbU8H3u5Rz56fH60jGFgr8uHwxs+aSKeqmluIVzM0m0kB7xQjKS6qPfd0b2ZoqQ==",
+      "license": "MIT",
+      "bin": {
+        "mustache": "bin/mustache"
+      }
+    },
     "node_modules/mute-stream": {
       "version": "2.0.0",
       "resolved": "https://registry.npmjs.org/mute-stream/-/mute-stream-2.0.0.tgz",
@@ -31127,7 +31296,6 @@
       "version": "0.13.1",
       "resolved": "https://registry.npmjs.org/retry/-/retry-0.13.1.tgz",
       "integrity": "sha512-XQBQ3I8W1Cge0Seh+6gjj03LbmRFWuoszgK9ooCpwYIrhhoO80pfq4cUkU5DkknwfOfFteRwlZ56PYOGYyFWdg==",
-      "dev": true,
       "license": "MIT",
       "engines": {
         "node": ">= 4"

package.json

@@ -85,6 +85,7 @@
     "@internationalized/number": "3.6.5",
     "@ionic/angular": "8.7.8",
     "@keyv/redis": "4.4.0",
+    "@langchain/core": "^0.3.80",
     "@nestjs/bull": "11.0.4",
     "@nestjs/cache-manager": "3.0.1",
     "@nestjs/common": "11.1.8",

tasks/tasks.md

@@ -209,6 +209,13 @@ Last updated: 2026-02-24
 - [x] Update AI and observability unit tests to assert LLM invocation trace behavior and keep provider fallback behavior stable.
 - [x] Run focused verification for touched AI suites and update task tracking notes.
 
+## Session Plan (2026-02-24, LangChain Wrapper + Arithmetic Direct Reply Correction)
+
+- [x] Enforce provider invocation through LangChain runnable wrapper in `AiService.generateText`.
+- [x] Fix no-tool arithmetic prompts to return computed deterministic replies instead of capability fallback text.
+- [x] Add/update unit tests for arithmetic direct replies and provider tracing/fallback behavior.
+- [x] Run focused verification (`test:ai` and `api:lint`) and update tracker notes.
+
 ## Verification Notes
 
 - `nx run api:lint` completed successfully (existing workspace warnings only).
@@ -282,3 +289,7 @@ Last updated: 2026-02-24
   - `npx nx run api:lint` (passes with existing workspace warnings)
   - `railway variable set -s ghostfolio-api --skip-deploys LANGCHAIN_API_KEY=... LANGSMITH_API_KEY=... LANGCHAIN_TRACING_V2=true LANGSMITH_TRACING=true LANGSMITH_PROJECT=ghostfolio-ai-agent`
   - `railway variable list -s ghostfolio-api --kv` confirms: `LANGCHAIN_API_KEY`, `LANGSMITH_API_KEY`, `LANGCHAIN_TRACING_V2`, `LANGSMITH_TRACING`, `LANGSMITH_PROJECT`
+- LangChain wrapper + arithmetic direct reply verification (local, 2026-02-24):
+  - `npx jest apps/api/src/app/endpoints/ai/ai-agent.utils.spec.ts apps/api/src/app/endpoints/ai/ai.service.spec.ts apps/api/src/app/endpoints/ai/ai-observability.service.spec.ts --config apps/api/jest.config.ts` (36/36 tests passed)
+  - `npm run test:ai` (9/9 suites passed, 49/49 tests)
+  - `npx nx run api:lint --verbose` (passes with existing workspace warnings)