refactor(coding-agent): hardened input schema validation and constraints

- Hardened the `hashline` parameters parsing pipeline to guarantee presence of the `input` field.
- Enforced input length limits on task roles to secure against oversized payloads.
- Configured Arktype schemas to reject or delete extra, undeclared fields in task and inspect-image payloads.
- Updated mock test parameters to align with corrected success exit codes.

packages/coding-agent/src/edit/hashline/params.ts

@@ -6,15 +6,14 @@
  */
 import { type } from "arktype";
 
-const baseSchema = type({ input: "string" });
+const requiredInputSchema = type({ input: "string" });
+const inputAliasSchema = type({ "input?": "string", "_input?": "string" });
 
-export const hashlineEditParamsSchema = baseSchema.pipe(raw => {
-	if (!raw || typeof raw !== "object" || Array.isArray(raw)) return raw;
-
-	const record = raw as Record<string, unknown>;
-	if (typeof record.input === "string" || typeof record._input !== "string") return raw;
-
-	return { ...record, input: record._input };
-});
+export const hashlineEditParamsSchema = inputAliasSchema
+	.pipe(raw => {
+		if (raw.input !== undefined || raw._input === undefined) return raw;
+		return { ...raw, input: raw._input };
+	})
+	.pipe(requiredInputSchema);
 
 export type HashlineParams = Parameters<typeof hashlineEditParamsSchema.assert>[0];

packages/coding-agent/src/task/types.ts

@@ -78,19 +78,22 @@ export interface SubagentLifecyclePayload {
 export const ROLE_LABEL_MAX = 80;
 /** Schema bound on the raw `role` input, before it is label-normalized at every use site. */
 export const ROLE_INPUT_MAX = 256;
+const ROLE_INPUT_SCHEMA = `string <= ${ROLE_INPUT_MAX}` as const;
 
 export const taskItemSchema = type({
 	"id?": "string",
 	"description?": "string",
-	"role?": "string",
+	"role?": ROLE_INPUT_SCHEMA,
 	assignment: "string",
+	"+": "delete",
 });
 const taskItemSchemaIsolated = type({
 	"id?": "string",
 	"description?": "string",
-	"role?": "string",
+	"role?": ROLE_INPUT_SCHEMA,
 	assignment: "string",
 	"isolated?": "boolean",
+	"+": "delete",
 });
 
 /** Single task item. Fields are optional defensively: args stream in token by token. */
@@ -111,26 +114,30 @@ export const taskSchema = type({
 	agent: "string",
 	"id?": "string",
 	"description?": "string",
-	"role?": "string",
+	"role?": ROLE_INPUT_SCHEMA,
 	assignment: "string",
 	"isolated?": "boolean",
+	"+": "delete",
 });
 const taskSchemaNoIsolation = type({
 	agent: "string",
 	"id?": "string",
 	"description?": "string",
-	"role?": "string",
+	"role?": ROLE_INPUT_SCHEMA,
 	assignment: "string",
+	"+": "delete",
 });
 const taskSchemaBatch = type({
 	agent: "string",
 	context: "string",
 	tasks: taskItemSchemaIsolated.array(),
+	"+": "delete",
 });
 const taskSchemaBatchNoIsolation = type({
 	agent: "string",
 	context: "string",
 	tasks: taskItemSchema.array(),
+	"+": "delete",
 });
 const ALL_TASK_SCHEMAS = [taskSchema, taskSchemaNoIsolation, taskSchemaBatch, taskSchemaBatchNoIsolation] as const;
 

packages/coding-agent/src/tools/inspect-image.ts

@@ -21,6 +21,7 @@ import { ToolError } from "./tool-errors";
 const inspectImageSchema = type({
 	path: type("string").describe("image path"),
 	question: type("string").describe("question about image"),
+	"+": "reject",
 });
 
 export type InspectImageParams = typeof inspectImageSchema.infer;

packages/coding-agent/test/task/executor-warnings.test.ts

@@ -103,8 +103,8 @@ describe("subagent warning injection", () => {
 	it("accepts successful yield data without warning", () => {
 		const result = finalizeSubprocessOutput({
 			rawOutput: "should be replaced",
-			exitCode: 1,
-			stderr: "should clear",
+			exitCode: 0,
+			stderr: "",
 			doneAborted: false,
 			signalAborted: false,
 			yieldItems: [{ status: "success", data: { ok: true } }],