feat(executor): per-task agent MCP server + token registry (Phase 2)

Adds the agent-facing MCP transport foundation: a Registry that mints a
per-task bearer token bound to a freshly built MCP server exposing the four
agent tools (ask_user, report_summary, spawn_subtask, record_progress), and
an HTTP handler (StreamableHTTP) that resolves the token to that server. The
server never trusts an agent-supplied task ID — context comes from the token.

The default storeChannel now buffers summary and question signals under a
mutex (an MCP tool call lands on an HTTP-handler goroutine mid-run), exposing
ReportedSummary/PendingQuestion. The pool flushes the buffered summary onto
the execution after the run, replacing the runner's direct exec.Summary write
and keeping the read race-free.

ask_user follows the record-and-resume model: it buffers the question, returns
ErrAgentBlocked, and the tool tells the agent to end its turn; the run blocks
and resumes later via claude --resume (no live slot held).

Tests cover registry lifecycle, in-memory tool dispatch, and HTTP end-to-end
with bearer auth (valid token dispatches; invalid token rejected).

Not yet wired into the runners or mounted on the API server — next increment.

https://claude.ai/code/session_01SESwn7kQ7oP62trWw6pc39

1 files changed, 6 insertions, 6 deletions

diff --git a/internal/executor/container_test.go b/internal/executor/container_test.go
index 5ee3a3c..9cd80dc 100644
--- a/internal/executor/container_test.go
+++ b/internal/executor/container_test.go
@@ -139,7 +139,7 @@ func TestContainerRunner_Run_PreservesWorkspaceOnFailure(t *testing.T) {
 	}
 	exec := &storage.Execution{ID: "test-exec", TaskID: "test-task"}
 
-	err := runner.Run(context.Background(), tk, exec, newStoreChannel(nil, tk.ID, exec))
+	err := runner.Run(context.Background(), tk, exec, newStoreChannel(nil, tk.ID))
 	if err == nil {
 		t.Fatal("expected error due to mocked docker failure")
 	}
@@ -378,7 +378,7 @@ func TestContainerRunner_MissingCredentials_FailsFast(t *testing.T) {
 	}
 	e := &storage.Execution{ID: "test-exec", TaskID: "test-missing-creds"}
 
-	err := runner.Run(context.Background(), tk, e, newStoreChannel(nil, tk.ID, e))
+	err := runner.Run(context.Background(), tk, e, newStoreChannel(nil, tk.ID))
 	if err == nil {
 		t.Fatal("expected error due to missing credentials, got nil")
 	}
@@ -418,7 +418,7 @@ func TestContainerRunner_MissingSettings_FailsFast(t *testing.T) {
 	}
 	e := &storage.Execution{ID: "test-exec-2", TaskID: "test-missing-settings"}
 
-	err := runner.Run(context.Background(), tk, e, newStoreChannel(nil, tk.ID, e))
+	err := runner.Run(context.Background(), tk, e, newStoreChannel(nil, tk.ID))
 	if err == nil {
 		t.Fatal("expected error due to missing settings, got nil")
 	}
@@ -504,7 +504,7 @@ func TestContainerRunner_AuthError_SyncsAndRetries(t *testing.T) {
 	e := &storage.Execution{ID: "auth-retry-exec", TaskID: "auth-retry-test"}
 
 	// Run — first attempt will fail with auth error, triggering sync+retry
-	runner.Run(context.Background(), tk, e, newStoreChannel(nil, tk.ID, e))
+	runner.Run(context.Background(), tk, e, newStoreChannel(nil, tk.ID))
 	// We don't check error strictly since second run may also fail (git push etc.)
 	// What we care about is that docker was called twice and sync was called
 	if callCount < 2 {
@@ -550,7 +550,7 @@ func TestContainerRunner_ClonesStoryBranch(t *testing.T) {
 	}
 	e := &storage.Execution{ID: "exec-1", TaskID: "story-branch-test"}
 
-	runner.Run(context.Background(), tk, e, newStoreChannel(nil, tk.ID, e))
+	runner.Run(context.Background(), tk, e, newStoreChannel(nil, tk.ID))
 	os.RemoveAll(e.SandboxDir)
 
 	// Assert git checkout was called with the story branch name.
@@ -597,7 +597,7 @@ func TestContainerRunner_ClonesDefaultBranchWhenNoBranchName(t *testing.T) {
 	}
 	e := &storage.Execution{ID: "exec-2", TaskID: "no-branch-test"}
 
-	runner.Run(context.Background(), tk, e, newStoreChannel(nil, tk.ID, e))
+	runner.Run(context.Background(), tk, e, newStoreChannel(nil, tk.ID))
 	os.RemoveAll(e.SandboxDir)
 
 	for _, a := range cloneArgs {