feat: gate lgtm on robot review acceptance

test/review-badges.test.ts

@@ -36,6 +36,7 @@ describe("getReviewBadges", () => {
           scope: "task evidence",
           observations: ["Observed one unchecked edge case"],
           blind_spots: "Did not inspect prod traffic",
+          accepted: false,
           evidence_complete: false,
           evidence_convincing: false,
           missing_evidence: ["Prod traffic sample"],

test/robot-review.test.ts

@@ -36,6 +36,7 @@ describe("robot review helpers", () => {
     expect(reviews).toHaveLength(1);
     expect(reviews[0].reviewer).toBe("opencode");
     expect(reviews[0].iteration).toBe(1);
+    expect(reviews[0].accepted).toBe(true);
   });
 
   it("appends robot reviews as iterations", () => {
@@ -45,6 +46,7 @@ describe("robot review helpers", () => {
       scope: "task evidence",
       observations: ["Observed missing benchmark output"],
       blind_spots: "Did not inspect prod config",
+      accepted: false,
       evidence_complete: false,
       evidence_convincing: false,
       missing_evidence: ["Benchmark output for the claimed speedup"],
@@ -57,6 +59,7 @@ describe("robot review helpers", () => {
       scope: "updated task evidence",
       observations: ["Observed benchmark output and test transcript"],
       blind_spots: "Did not inspect long-run stability",
+      accepted: true,
       evidence_complete: true,
       evidence_convincing: true,
       missing_evidence: [],