fix(tools): flatten nested anyOf schemas for Vertex AI compatibility

Claude API on Vertex AI (Cloud Code Assist) rejects nested anyOf schemas as invalid JSON Schema draft 2020-12. This change: - Add tryFlattenLiteralAnyOf() to convert Type.Union([Type.Literal(...)]) patterns from anyOf with const values to flat enum arrays - Update stringEnum helper in bash-tools to use Type.Unsafe with flat enum - Flatten BrowserActSchema from discriminated union to single object - Simplify TelegramToolSchema to use Type.String() for IDs Fixes 400 errors when sending messages through WhatsApp/Telegram providers.
2026-01-07 16:54:13 +00:00
parent de55f4e111
commit a2b3f2c18a
4 changed files with 132 additions and 80 deletions
--- a/src/agents/bash-tools.ts
+++ b/src/agents/bash-tools.ts
@@ -39,17 +39,19 @@ const DEFAULT_PATH =
  process.env.PATH ??
  "/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin";

-const stringEnum = (
-  values: readonly string[],
-  options?: Parameters<typeof Type.Union>[1],
+// NOTE: Using Type.Unsafe with enum instead of Type.Union([Type.Literal(...)])
+// because Claude API on Vertex AI rejects nested anyOf schemas as invalid JSON Schema.
+// Type.Union of literals compiles to { anyOf: [{enum:["a"]}, {enum:["b"]}, ...] }
+// which is valid but not accepted. A flat enum { type: "string", enum: [...] } works.
+const stringEnum = <T extends readonly string[]>(
+  values: T,
+  options?: { description?: string },
 ) =>
-  Type.Union(
-    values.map((value) => Type.Literal(value)) as [
-      ReturnType<typeof Type.Literal>,
-      ...ReturnType<typeof Type.Literal>[],
-    ],
-    options,
-  );
+  Type.Unsafe<T[number]>({
+    type: "string",
+    enum: values as unknown as string[],
+    ...options,
+  });

 export type BashToolDefaults = {
  backgroundMs?: number;
--- a/src/agents/pi-tools.ts
+++ b/src/agents/pi-tools.ts
@@ -154,12 +154,73 @@ function mergePropertySchemas(existing: unknown, incoming: unknown): unknown {
  return existing;
 }

+// Check if an anyOf array contains only literal values that can be flattened
+// TypeBox Type.Literal generates { const: "value", type: "string" }
+// Some schemas may use { enum: ["value"], type: "string" }
+// Both patterns are flattened to { type: "string", enum: ["a", "b", ...] }
+function tryFlattenLiteralAnyOf(
+  anyOf: unknown[],
+): { type: string; enum: unknown[] } | null {
+  if (anyOf.length === 0) return null;
+
+  const allValues: unknown[] = [];
+  let commonType: string | null = null;
+
+  for (const variant of anyOf) {
+    if (!variant || typeof variant !== "object") return null;
+    const v = variant as Record<string, unknown>;
+
+    // Extract the literal value - either from const or single-element enum
+    let literalValue: unknown;
+    if ("const" in v) {
+      literalValue = v.const;
+    } else if (Array.isArray(v.enum) && v.enum.length === 1) {
+      literalValue = v.enum[0];
+    } else {
+      return null; // Not a literal pattern
+    }
+
+    // Must have consistent type (usually "string")
+    const variantType = typeof v.type === "string" ? v.type : null;
+    if (!variantType) return null;
+    if (commonType === null) commonType = variantType;
+    else if (commonType !== variantType) return null;
+
+    allValues.push(literalValue);
+  }
+
+  if (commonType && allValues.length > 0) {
+    return { type: commonType, enum: allValues };
+  }
+  return null;
+}
+
 function cleanSchemaForGemini(schema: unknown): unknown {
  if (!schema || typeof schema !== "object") return schema;
  if (Array.isArray(schema)) return schema.map(cleanSchemaForGemini);

  const obj = schema as Record<string, unknown>;
  const hasAnyOf = "anyOf" in obj && Array.isArray(obj.anyOf);
+
+  // Try to flatten anyOf of literals to a single enum BEFORE processing
+  // This handles Type.Union([Type.Literal("a"), Type.Literal("b")]) patterns
+  if (hasAnyOf) {
+    const flattened = tryFlattenLiteralAnyOf(obj.anyOf as unknown[]);
+    if (flattened) {
+      // Return flattened enum, preserving metadata (description, title, default, examples)
+      const result: Record<string, unknown> = {
+        type: flattened.type,
+        enum: flattened.enum,
+      };
+      for (const key of ["description", "title", "default", "examples"]) {
+        if (key in obj && obj[key] !== undefined) {
+          result[key] = obj[key];
+        }
+      }
+      return result;
+    }
+  }
+
  const cleaned: Record<string, unknown> = {};

  for (const [key, value] of Object.entries(obj)) {
@@ -409,8 +470,13 @@ function createWhatsAppLoginTool(): AnyAgentTool {
    name: "whatsapp_login",
    description:
      "Generate a WhatsApp QR code for linking, or wait for the scan to complete.",
+    // NOTE: Using Type.Unsafe for action enum instead of Type.Union([Type.Literal(...)])
+    // because Claude API on Vertex AI rejects nested anyOf schemas as invalid JSON Schema.
    parameters: Type.Object({
-      action: Type.Union([Type.Literal("start"), Type.Literal("wait")]),
+      action: Type.Unsafe<"start" | "wait">({
+        type: "string",
+        enum: ["start", "wait"],
+      }),
      timeoutMs: Type.Optional(Type.Number()),
      force: Type.Optional(Type.Boolean()),
    }),
--- a/src/agents/tools/browser-tool.ts
+++ b/src/agents/tools/browser-tool.ts
@@ -28,74 +28,55 @@ import {
  readStringParam,
 } from "./common.js";

-const BrowserActSchema = Type.Union([
-  Type.Object({
-    kind: Type.Literal("click"),
-    ref: Type.String(),
-    targetId: Type.Optional(Type.String()),
-    doubleClick: Type.Optional(Type.Boolean()),
-    button: Type.Optional(Type.String()),
-    modifiers: Type.Optional(Type.Array(Type.String())),
+// NOTE: Using a flattened object schema instead of Type.Union([Type.Object(...), ...])
+// because Claude API on Vertex AI rejects nested anyOf schemas as invalid JSON Schema.
+// The discriminator (kind) determines which properties are relevant; runtime validates.
+const BrowserActSchema = Type.Object({
+  kind: Type.Unsafe<string>({
+    type: "string",
+    enum: [
+      "click",
+      "type",
+      "press",
+      "hover",
+      "drag",
+      "select",
+      "fill",
+      "resize",
+      "wait",
+      "evaluate",
+      "close",
+    ],
  }),
-  Type.Object({
-    kind: Type.Literal("type"),
-    ref: Type.String(),
-    text: Type.String(),
-    targetId: Type.Optional(Type.String()),
-    submit: Type.Optional(Type.Boolean()),
-    slowly: Type.Optional(Type.Boolean()),
-  }),
-  Type.Object({
-    kind: Type.Literal("press"),
-    key: Type.String(),
-    targetId: Type.Optional(Type.String()),
-  }),
-  Type.Object({
-    kind: Type.Literal("hover"),
-    ref: Type.String(),
-    targetId: Type.Optional(Type.String()),
-  }),
-  Type.Object({
-    kind: Type.Literal("drag"),
-    startRef: Type.String(),
-    endRef: Type.String(),
-    targetId: Type.Optional(Type.String()),
-  }),
-  Type.Object({
-    kind: Type.Literal("select"),
-    ref: Type.String(),
-    values: Type.Array(Type.String()),
-    targetId: Type.Optional(Type.String()),
-  }),
-  Type.Object({
-    kind: Type.Literal("fill"),
-    fields: Type.Array(Type.Record(Type.String(), Type.Unknown())),
-    targetId: Type.Optional(Type.String()),
-  }),
-  Type.Object({
-    kind: Type.Literal("resize"),
-    width: Type.Number(),
-    height: Type.Number(),
-    targetId: Type.Optional(Type.String()),
-  }),
-  Type.Object({
-    kind: Type.Literal("wait"),
-    timeMs: Type.Optional(Type.Number()),
-    text: Type.Optional(Type.String()),
-    textGone: Type.Optional(Type.String()),
-    targetId: Type.Optional(Type.String()),
-  }),
-  Type.Object({
-    kind: Type.Literal("evaluate"),
-    fn: Type.String(),
-    ref: Type.Optional(Type.String()),
-    targetId: Type.Optional(Type.String()),
-  }),
-  Type.Object({
-    kind: Type.Literal("close"),
-    targetId: Type.Optional(Type.String()),
-  }),
-]);
+  // Common fields
+  targetId: Type.Optional(Type.String()),
+  ref: Type.Optional(Type.String()),
+  // click
+  doubleClick: Type.Optional(Type.Boolean()),
+  button: Type.Optional(Type.String()),
+  modifiers: Type.Optional(Type.Array(Type.String())),
+  // type
+  text: Type.Optional(Type.String()),
+  submit: Type.Optional(Type.Boolean()),
+  slowly: Type.Optional(Type.Boolean()),
+  // press
+  key: Type.Optional(Type.String()),
+  // drag
+  startRef: Type.Optional(Type.String()),
+  endRef: Type.Optional(Type.String()),
+  // select
+  values: Type.Optional(Type.Array(Type.String())),
+  // fill - use permissive array of objects
+  fields: Type.Optional(Type.Array(Type.Object({}, { additionalProperties: true }))),
+  // resize
+  width: Type.Optional(Type.Number()),
+  height: Type.Optional(Type.Number()),
+  // wait
+  timeMs: Type.Optional(Type.Number()),
+  textGone: Type.Optional(Type.String()),
+  // evaluate
+  fn: Type.Optional(Type.String()),
+});

 // IMPORTANT: OpenAI function tool schemas must have a top-level `type: "object"`.
 // A root-level `Type.Union([...])` compiles to `{ anyOf: [...] }` (no `type`),
--- a/src/agents/tools/telegram-schema.ts
+++ b/src/agents/tools/telegram-schema.ts
@@ -2,11 +2,14 @@ import { Type } from "@sinclair/typebox";

 import { createReactionSchema } from "./reaction-schema.js";

+// NOTE: chatId and messageId use Type.String() instead of Type.Union([Type.String(), Type.Number()])
+// because nested anyOf schemas cause JSON Schema validation failures with Claude API on Vertex AI.
+// Telegram IDs are coerced to strings at runtime in telegram-actions.ts.
 export const TelegramToolSchema = Type.Union([
  createReactionSchema({
    ids: {
-      chatId: Type.Union([Type.String(), Type.Number()]),
-      messageId: Type.Union([Type.String(), Type.Number()]),
+      chatId: Type.String(),
+      messageId: Type.String(),
    },
    includeRemove: true,
  }),