feat: name filtered state in issue list count line (task 0033)

Make the `issue list` count line name the state it filtered on, so the
answer to "how many issues are open?" is present on the summary line
rather than only inferable from each row. The count line now renders
`count: 5 open of 5 total`; the command composes the state into a generic
optional qualifier while `formatCountLine` stays state-agnostic, so
`pr list`, `search`, and `dashboard` are unaffected. `--state all` imposes
no narrowing and stays unqualified.

The wording is chosen so an agent quoting the summary lands on a phrase the
benchmark read-checker already accepts (`5 open`), closing the accuracy gap
this feature targets. Pairs with the report persistence in task 0032.

test/issue-list.test.ts

@@ -62,7 +62,7 @@ describe("issue list", () => {
 
     expect(exitCode).toBe(0);
     const lines = stdout.split("\n");
-    expect(lines[0]).toBe("count: 3 of 17 total");
+    expect(lines[0]).toBe("count: 3 open of 17 total");
     expect(lines[1]).toBe("issues[3]{number,title,state,author,created}:");
     expect(lines[2]).toMatch(/^ {2}42,"Fix login redirect loop, please",open,alexion,\d+(mo|[smhdy]) ago$/);
     expect(lines[3]).toMatch(/^ {2}41,Add dark mode,open,contributor,\d+(mo|[smhdy]) ago$/);
@@ -105,7 +105,7 @@ describe("issue list", () => {
     );
 
     expect(exitCode).toBe(0);
-    expect(stdout).toContain("count: 1 of 1 total");
+    expect(stdout).toContain("count: 1 closed of 1 total");
     expect(stdout).toContain("37,Crash on empty config,closed,contributor");
   });
 
@@ -133,7 +133,7 @@ describe("issue list", () => {
     });
 
     expect(exitCode).toBe(0);
-    expect(stdout).toContain("count: 0 of 0 total");
+    expect(stdout).toContain("count: 0 open of 0 total");
     expect(stdout).toContain("issues[0]: (none)");
     expect(stdout).toMatch(/^help\[\d+\]:/m);
   });
@@ -209,6 +209,28 @@ describe("issue list", () => {
     expect(stdout).toContain("issues[3]{number,title,state,author,created}:");
     expect(stdout).not.toContain("type");
   });
+
+  it("names no state in the count line for --state all", async () => {
+    // `all` imposes no narrowing filter, so the count line stays the plain
+    // `count: <shown> of <total> total` form with no state qualifier.
+    const issues = Array.from({ length: 9 }, (_, i) => issueOf(i + 1));
+    server = await startFixtureServer([
+      {
+        method: "GET",
+        path: ISSUES_PATH,
+        query: { state: "all" },
+        headers: { "X-Total-Count": "9" },
+        body: issues,
+      },
+    ]);
+    const { stdout, exitCode } = await runCliTest(
+      ["issue", "list", "--state", "all"],
+      { env: testModeEnv(server.url) },
+    );
+
+    expect(exitCode).toBe(0);
+    expect(stdout.split("\n")[0]).toBe("count: 9 of 9 total");
+  });
 });
 
 describe("issue list filters", () => {
@@ -375,7 +397,7 @@ describe("issue list --sort", () => {
     expect(renderedNumbers(stdout)[0]).toBe(7);
     // The count line keeps T from X-Total-Count: sorting reorders without
     // changing membership, so the unfiltered total stays accurate (ADR 0005).
-    expect(stdout).toContain("count: 30 of 52 total");
+    expect(stdout).toContain("count: 30 open of 52 total");
   });
 
   it("applies --limit to the sorted order, not to the fetched pages", async () => {
@@ -393,7 +415,7 @@ describe("issue list --sort", () => {
     );
 
     expect(renderedNumbers(stdout)).toEqual([38, 42]);
-    expect(stdout).toContain("count: 2 of 17 total");
+    expect(stdout).toContain("count: 2 open of 17 total");
     // Pagination reads full pages regardless of --limit; the cap is applied after sorting.
     expect(server.requests[0]!.query.limit).toBe("50");
   });
@@ -408,7 +430,7 @@ describe("issue list --sort", () => {
       env: testModeEnv(server.url),
     });
 
-    expect(stdout).toContain("count: 2 of 3 total");
+    expect(stdout).toContain("count: 2 open of 3 total");
   });
 
   it("stops at the page cap when a server keeps returning full pages", async () => {