Skip to content

YC (S26) | Open Computer History | Continuously record your company computer work, map your workflows, help you find work worth automating, and power your agents' context

fix(e2e): assert the ACP permission heading the runtime actually sends

Louis Beaumont92415f854e40
diff --git a/apps/screenpipe-app-tauri/e2e/specs/acp-backend.spec.ts b/apps/screenpipe-app-tauri/e2e/specs/acp-backend.spec.ts
index 688355ad9..262c5628c 100644
--- a/apps/screenpipe-app-tauri/e2e/specs/acp-backend.spec.ts
+++ b/apps/screenpipe-app-tauri/e2e/specs/acp-backend.spec.ts
@@ -617,7 +617,11 @@ describe("ACP backend", function () {
const permission = beforeApproval.find(
(envelope) => envelope.event?.type === "extension_ui_request",
)?.event;
- expect(permission?.title).toContain("acp:permission:Write mock result");
+ // The heading comes from the tool's `kind` (permission_label), not its raw
+ // title, so a shell command can never be humanized into a mangled heading.
+ // The mock's tool is kind "edit", so the heading is "Edit a file" and the
+ // raw target ("Write mock result") is carried verbatim in the detail below.
+ expect(permission?.title).toContain("acp:permission:Edit a file");
expect(permission?.options.map((option: any) => option.kind)).toEqual([
"allow_once",
"reject_once",
@@ -625,6 +629,9 @@ describe("ACP backend", function () {
await foregroundChat(normalSession);
const permissionCard = await waitForAgentAction("permission");
+ // Heading plus the verbatim target: losing either would leave the user
+ // approving a write without seeing what is being written.
+ expect(await permissionCard.getText()).toContain("Edit a file");
expect(await permissionCard.getText()).toContain("Write mock result");
expect(await permissionCard.getText()).toContain("Allow once");
await answerAgentAction("permission", "Allow once");