fix(ai): corrected AI tool handling via JSON-schema validation flow

- Replaced fromTypeBox conversion with a JSON-schema validator flow in ai tool handling and execution paths.
- Added recursive schema validation and expanded TypeBox checks for refs, enums, uniqueItems, and constraint keywords.
- Sanitized Azure/CCA tool schemas by dropping unsupported fields and rewriting oneOf tool branches as anyOf.
- Tightened argument and model-config validation, preserving unknown tool fields and adding apiKey plus compatibility flags.

packages/coding-agent/test/commit-shared-llm.test.ts

@@ -0,0 +1,33 @@
+import { describe, expect, it } from "bun:test";
+import type { AssistantMessage } from "@oh-my-pi/pi-ai";
+import { createConventionalAnalysisTool, parseConventionalAnalysisResponse } from "../src/commit/shared-llm";
+
+describe("commit shared LLM parsing", () => {
+	it("ignores harmless extra fields in conventional analysis tool output", () => {
+		const tool = createConventionalAnalysisTool("Analyze a diff.");
+		const message = {
+			role: "assistant",
+			content: [
+				{
+					type: "toolCall",
+					id: "call-analysis",
+					name: tool.name,
+					arguments: {
+						type: "fix",
+						scope: null,
+						details: [],
+						issue_refs: [],
+						summary: "fix: handle parser edge case",
+					},
+				},
+			],
+		} as unknown as AssistantMessage;
+
+		expect(parseConventionalAnalysisResponse(message, tool)).toEqual({
+			type: "fix",
+			scope: null,
+			details: [],
+			issueRefs: [],
+		});
+	});
+});