fix(agents): normalize requester peer id before bound-account lookup

Delivery targets on Matrix (and other channels that namespace the `to`
field) arrive in `kind:<id>` form — for example `room:!abc:example.org`
— while route bindings store the raw peer id on `match.peer.id`
(`!abc:example.org`). Passing `ctx.agentTo` directly to
`resolveFirstBoundAccountId` caused exact peer matches to silently fail
and the lookup to fall through to channel-only or caller-account
fallback, so cross-agent spawns could still post as the wrong identity
when only a peer-specific binding was configured.

- src/agents/subagent-spawn.ts: strip known delivery-target prefixes
  (`room:`, `channel:`, `chat:`, `user:`, `dm:`, `group:`, and the
  channel-namespaced `${channelId}:`) from `requesterTo` before handing
  it to `resolveFirstBoundAccountId`. The inferred `peerKind` still uses
  the original `requesterTo` so channel plugins can apply their own
  inference on the wire format.

Regression test in lifecycle suite:
- sessions_spawn strips channel-side prefixes from agentTo before the
  bound-account lookup — a binding on the raw room id resolves correctly
  even when the caller's `agentTo` is `room:<id>`.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

src/agents/openclaw-tools.subagents.sessions-spawn.lifecycle.test.ts

@@ -652,6 +652,51 @@ describe("openclaw-tools: subagents (sessions_spawn lifecycle)", () => {
     expect(spawnAccountId).toBe("bot-alpha-room-a");
   });
 
+  it("sessions_spawn strips channel-side prefixes from agentTo before bound-account lookup", async () => {
+    let spawnAccountId: string | undefined;
+    const rawRoomId = "!exampleRoomId:example.org";
+    setSessionsSpawnConfigOverride({
+      session: { mainKey: "main", scope: "per-sender" },
+      messages: { queue: { debounceMs: 0 } },
+      agents: { defaults: { subagents: { allowAgents: ["bot-alpha"] } } },
+      bindings: [
+        {
+          type: "route",
+          agentId: "bot-alpha",
+          match: {
+            channel: "matrix",
+            peer: { kind: "channel", id: rawRoomId },
+            accountId: "bot-alpha",
+          },
+        },
+      ],
+    });
+    setupSessionsSpawnGatewayMock({
+      onAgentSubagentSpawn: (params) => {
+        const rec = params as { accountId?: string } | undefined;
+        spawnAccountId = rec?.accountId;
+      },
+    });
+
+    // agentTo arrives in delivery-target format (room:<id>), while the binding
+    // stores the raw id. Without prefix normalization the exact peer match
+    // would silently fail and the caller account would leak to the child.
+    const tool = await getSessionsSpawnTool({
+      agentSessionKey: "main",
+      agentChannel: "matrix",
+      agentAccountId: "bot-beta",
+      agentTo: `room:${rawRoomId}`,
+    });
+
+    const result = await tool.execute("call-prefixed-to", {
+      task: "do thing",
+      agentId: "bot-alpha",
+      cleanup: "keep",
+    });
+    expect(result.details).toMatchObject({ status: "accepted", runId: expect.any(String) });
+    expect(spawnAccountId).toBe("bot-alpha");
+  });
+
   it("sessions_spawn preserves the caller's account for same-agent subagent spawns", async () => {
     let spawnAccountId: string | undefined;
     const room = "!someRoom:example.org";

src/agents/subagent-spawn.ts

@@ -287,6 +287,10 @@ function summarizeError(err: unknown): string {
   return "error";
 }
 
+// Delivery targets carry a channel-side prefix (e.g. Matrix uses `room:<roomId>`;
+// LINE uses `line:group:<id>`), but route bindings store raw peer ids on
+// `match.peer.id`. Peel the `<channel>:` namespace first, then loop over generic
+// target-kind prefixes so the raw peer id surfaces.
 const KIND_PREFIX_TO_CHAT_TYPE: Readonly<Record<string, ChatType>> = {
   "room:": "channel",
   "channel:": "channel",