opencode-swarm-plugin 0.26.1 → 0.27.0

package/src/swarm-review.test.ts

@@ -0,0 +1,702 @@
+/**
+ * Swarm Structured Review Tests
+ *
+ * Tests for the coordinator-driven review of worker output.
+ * The review is epic-aware - it checks if work serves the overall goal
+ * and enables downstream tasks.
+ *
+ * Credit: Review patterns inspired by https://github.com/nexxeln/opencode-config
+ */
+import { describe, it, expect, beforeEach, vi } from "vitest";
+import {
+  generateReviewPrompt,
+  ReviewResultSchema,
+  markReviewApproved,
+  isReviewApproved,
+  getReviewStatus,
+  clearReviewStatus,
+  swarm_review,
+  swarm_review_feedback,
+  type ReviewPromptContext,
+  type ReviewResult,
+  type ReviewIssue,
+} from "./swarm-review";
+
+// Mock swarm-mail
+vi.mock("swarm-mail", () => ({
+  sendSwarmMessage: vi.fn().mockResolvedValue({ success: true }),
+}));
+
+const mockContext = {
+  sessionID: `test-review-${Date.now()}`,
+  messageID: `test-message-${Date.now()}`,
+  agent: "test-agent",
+  abort: new AbortController().signal,
+};
+
+// ============================================================================
+// Review Prompt Generation
+// ============================================================================
+
+describe("generateReviewPrompt", () => {
+  const baseContext: ReviewPromptContext = {
+    epic_id: "bd-test-123",
+    epic_title: "Add user authentication",
+    epic_description: "Implement OAuth2 with JWT tokens",
+    task_id: "bd-test-123.1",
+    task_title: "Create auth utilities",
+    task_description: "JWT sign/verify functions",
+    files_touched: ["src/lib/auth.ts"],
+    diff: "+export function signToken() {}",
+  };
+
+  it("includes epic goal for big-picture context", () => {
+    const prompt = generateReviewPrompt(baseContext);
+    expect(prompt).toContain("Add user authentication");
+    expect(prompt).toContain("OAuth2 with JWT tokens");
+    expect(prompt).toContain("## Epic Goal");
+  });
+
+  it("includes task requirements", () => {
+    const prompt = generateReviewPrompt(baseContext);
+    expect(prompt).toContain("Create auth utilities");
+    expect(prompt).toContain("JWT sign/verify functions");
+    expect(prompt).toContain("## Task Requirements");
+  });
+
+  it("includes dependency context (what this builds on)", () => {
+    const contextWithDeps: ReviewPromptContext = {
+      ...baseContext,
+      task_id: "bd-test-123.2",
+      task_title: "Create auth middleware",
+      completed_dependencies: [
+        {
+          id: "bd-test-123.1",
+          title: "Create auth utilities",
+          summary: "JWT sign/verify done",
+        },
+      ],
+    };
+    const prompt = generateReviewPrompt(contextWithDeps);
+    expect(prompt).toContain("This Task Builds On");
+    expect(prompt).toContain("Create auth utilities");
+    expect(prompt).toContain("JWT sign/verify done");
+  });
+
+  it("includes downstream context (what depends on this)", () => {
+    const contextWithDownstream: ReviewPromptContext = {
+      ...baseContext,
+      downstream_tasks: [
+        { id: "bd-test-123.2", title: "Create auth middleware" },
+        { id: "bd-test-123.3", title: "Add protected routes" },
+      ],
+    };
+    const prompt = generateReviewPrompt(contextWithDownstream);
+    expect(prompt).toContain("Downstream Tasks");
+    expect(prompt).toContain("Create auth middleware");
+    expect(prompt).toContain("Add protected routes");
+  });
+
+  it("includes the actual code diff", () => {
+    const diff = `+export function signToken(payload: TokenPayload): string {
++  return jwt.sign(payload, SECRET, { expiresIn: '1h' });
++}`;
+    const contextWithDiff: ReviewPromptContext = {
+      ...baseContext,
+      diff,
+    };
+    const prompt = generateReviewPrompt(contextWithDiff);
+    expect(prompt).toContain("signToken");
+    expect(prompt).toContain("TokenPayload");
+    expect(prompt).toContain("```diff");
+  });
+
+  it("includes review criteria checklist", () => {
+    const prompt = generateReviewPrompt(baseContext);
+    expect(prompt).toContain("Fulfills Requirements");
+    expect(prompt).toContain("Serves Epic Goal");
+    expect(prompt).toContain("Enables Downstream");
+    expect(prompt).toContain("Type Safety");
+    expect(prompt).toContain("No Critical Bugs");
+    expect(prompt).toContain("Test Coverage");
+  });
+
+  it("includes files modified section", () => {
+    const prompt = generateReviewPrompt(baseContext);
+    expect(prompt).toContain("## Files Modified");
+    expect(prompt).toContain("`src/lib/auth.ts`");
+  });
+
+  it("includes response format instructions", () => {
+    const prompt = generateReviewPrompt(baseContext);
+    expect(prompt).toContain("## Response Format");
+    expect(prompt).toContain('"status"');
+    expect(prompt).toContain('"approved"');
+    expect(prompt).toContain('"needs_changes"');
+  });
+});
+
+// ============================================================================
+// Review Result Schema
+// ============================================================================
+
+describe("ReviewResultSchema", () => {
+  it("accepts approved status with summary", () => {
+    const result: ReviewResult = {
+      status: "approved",
+      summary: "Clean implementation, exports are clear for downstream tasks",
+    };
+    expect(ReviewResultSchema.safeParse(result).success).toBe(true);
+  });
+
+  it("accepts needs_changes status with issues array", () => {
+    const result: ReviewResult = {
+      status: "needs_changes",
+      issues: [
+        {
+          file: "src/lib/auth.ts",
+          line: 42,
+          issue: "Missing error handling for expired tokens",
+          suggestion:
+            "Return { valid: false, error: 'expired' } instead of throwing",
+        },
+      ],
+      remaining_attempts: 2,
+    };
+    expect(ReviewResultSchema.safeParse(result).success).toBe(true);
+  });
+
+  it("requires issues array when status is needs_changes", () => {
+    const result = {
+      status: "needs_changes",
+      // missing issues array
+    };
+    const parsed = ReviewResultSchema.safeParse(result);
+    expect(parsed.success).toBe(false);
+  });
+
+  it("rejects needs_changes with empty issues array", () => {
+    const result = {
+      status: "needs_changes",
+      issues: [],
+    };
+    const parsed = ReviewResultSchema.safeParse(result);
+    expect(parsed.success).toBe(false);
+  });
+
+  it("tracks remaining review attempts", () => {
+    const result: ReviewResult = {
+      status: "needs_changes",
+      issues: [{ file: "x.ts", issue: "bug" }],
+      remaining_attempts: 1,
+    };
+    const parsed = ReviewResultSchema.safeParse(result);
+    expect(parsed.success).toBe(true);
+    if (parsed.success) {
+      expect(parsed.data.remaining_attempts).toBe(1);
+    }
+  });
+
+  it("accepts approved without summary", () => {
+    const result: ReviewResult = {
+      status: "approved",
+    };
+    expect(ReviewResultSchema.safeParse(result).success).toBe(true);
+  });
+
+  it("accepts issue without line number", () => {
+    const result: ReviewResult = {
+      status: "needs_changes",
+      issues: [{ file: "x.ts", issue: "general problem" }],
+    };
+    expect(ReviewResultSchema.safeParse(result).success).toBe(true);
+  });
+
+  it("accepts issue without suggestion", () => {
+    const result: ReviewResult = {
+      status: "needs_changes",
+      issues: [{ file: "x.ts", line: 10, issue: "problem here" }],
+    };
+    expect(ReviewResultSchema.safeParse(result).success).toBe(true);
+  });
+});
+
+// ============================================================================
+// Review Status Tracking
+// ============================================================================
+
+describe("Review status tracking", () => {
+  beforeEach(() => {
+    clearReviewStatus("test-task-1");
+    clearReviewStatus("test-task-2");
+  });
+
+  it("starts with no review status", () => {
+    const status = getReviewStatus("test-task-1");
+    expect(status.reviewed).toBe(false);
+    expect(status.approved).toBe(false);
+    expect(status.attempt_count).toBe(0);
+    expect(status.remaining_attempts).toBe(3);
+  });
+
+  it("marks task as approved", () => {
+    markReviewApproved("test-task-1");
+    expect(isReviewApproved("test-task-1")).toBe(true);
+    const status = getReviewStatus("test-task-1");
+    expect(status.reviewed).toBe(true);
+    expect(status.approved).toBe(true);
+  });
+
+  it("tracks separate status per task", () => {
+    markReviewApproved("test-task-1");
+    expect(isReviewApproved("test-task-1")).toBe(true);
+    expect(isReviewApproved("test-task-2")).toBe(false);
+  });
+
+  it("clears review status", () => {
+    markReviewApproved("test-task-1");
+    expect(isReviewApproved("test-task-1")).toBe(true);
+    clearReviewStatus("test-task-1");
+    expect(isReviewApproved("test-task-1")).toBe(false);
+  });
+});
+
+// ============================================================================
+// swarm_review tool
+// ============================================================================
+
+describe("swarm_review", () => {
+  it("has correct tool metadata", () => {
+    expect(swarm_review.description).toContain("review prompt");
+    expect(swarm_review.description).toContain("epic context");
+  });
+
+  it("returns JSON with review_prompt field", async () => {
+    // This test exercises the tool structure without needing real git/beads
+    // The tool will fail to get real data but should still return valid JSON
+    // Use /tmp which exists on all systems
+    const result = await swarm_review.execute(
+      {
+        project_key: "/tmp",
+        epic_id: "bd-test-123",
+        task_id: "bd-test-123.1",
+        files_touched: ["src/test.ts"],
+      },
+      mockContext
+    );
+
+    const parsed = JSON.parse(result);
+    expect(parsed).toHaveProperty("review_prompt");
+    expect(parsed).toHaveProperty("context");
+    expect(parsed.context.epic_id).toBe("bd-test-123");
+    expect(parsed.context.task_id).toBe("bd-test-123.1");
+  });
+
+  it("includes remaining attempts in context", async () => {
+    clearReviewStatus("bd-test-123.1");
+    const result = await swarm_review.execute(
+      {
+        project_key: "/tmp",
+        epic_id: "bd-test-123",
+        task_id: "bd-test-123.1",
+      },
+      mockContext
+    );
+
+    const parsed = JSON.parse(result);
+    expect(parsed.context.remaining_attempts).toBe(3);
+  });
+});
+
+// ============================================================================
+// swarm_review_feedback tool
+// ============================================================================
+
+describe("swarm_review_feedback", () => {
+  beforeEach(() => {
+    clearReviewStatus("bd-feedback-test");
+    vi.clearAllMocks();
+  });
+
+  it("has correct tool metadata", () => {
+    expect(swarm_review_feedback.description).toContain("feedback");
+    expect(swarm_review_feedback.description).toContain("max 3");
+  });
+
+  it("sends approved feedback successfully", async () => {
+    const result = await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test-project",
+        task_id: "bd-feedback-test",
+        worker_id: "worker-test",
+        status: "approved",
+        summary: "Looks good, clean implementation",
+      },
+      mockContext
+    );
+
+    const parsed = JSON.parse(result);
+    expect(parsed.success).toBe(true);
+    expect(parsed.status).toBe("approved");
+  });
+
+  it("sends needs_changes feedback with structured issues", async () => {
+    const issues: ReviewIssue[] = [
+      {
+        file: "src/auth.ts",
+        line: 42,
+        issue: "Missing null check",
+        suggestion: "Add if (!token) return null",
+      },
+    ];
+
+    const result = await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test-project",
+        task_id: "bd-feedback-test",
+        worker_id: "worker-test",
+        status: "needs_changes",
+        issues: JSON.stringify(issues),
+      },
+      mockContext
+    );
+
+    const parsed = JSON.parse(result);
+    expect(parsed.success).toBe(true);
+    expect(parsed.status).toBe("needs_changes");
+    expect(parsed.remaining_attempts).toBe(2);
+  });
+
+  it("requires issues for needs_changes status", async () => {
+    const result = await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test-project",
+        task_id: "bd-feedback-test",
+        worker_id: "worker-test",
+        status: "needs_changes",
+        // no issues provided
+      },
+      mockContext
+    );
+
+    const parsed = JSON.parse(result);
+    expect(parsed.success).toBe(false);
+    expect(parsed.error).toContain("requires at least one issue");
+  });
+
+  it("tracks review attempts (max 3)", async () => {
+    const issues = JSON.stringify([{ file: "x.ts", issue: "bug" }]);
+
+    // First attempt
+    let result = await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test-project",
+        task_id: "bd-feedback-test",
+        worker_id: "worker-test",
+        status: "needs_changes",
+        issues,
+      },
+      mockContext
+    );
+    let parsed = JSON.parse(result);
+    expect(parsed.attempt).toBe(1);
+    expect(parsed.remaining_attempts).toBe(2);
+
+    // Second attempt
+    result = await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test-project",
+        task_id: "bd-feedback-test",
+        worker_id: "worker-test",
+        status: "needs_changes",
+        issues,
+      },
+      mockContext
+    );
+    parsed = JSON.parse(result);
+    expect(parsed.attempt).toBe(2);
+    expect(parsed.remaining_attempts).toBe(1);
+  });
+
+  it("fails task after 3 rejected reviews", async () => {
+    const issues = JSON.stringify([{ file: "x.ts", issue: "still broken" }]);
+
+    // Exhaust all attempts
+    for (let i = 0; i < 3; i++) {
+      await swarm_review_feedback.execute(
+        {
+          project_key: "/tmp/test-project",
+          task_id: "bd-feedback-test",
+          worker_id: "worker-test",
+          status: "needs_changes",
+          issues,
+        },
+        mockContext
+      );
+    }
+
+    // Check final state
+    const status = getReviewStatus("bd-feedback-test");
+    expect(status.remaining_attempts).toBe(0);
+  });
+
+  it("clears attempts on approval", async () => {
+    const issues = JSON.stringify([{ file: "x.ts", issue: "bug" }]);
+
+    // Add some attempts
+    await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test-project",
+        task_id: "bd-feedback-test",
+        worker_id: "worker-test",
+        status: "needs_changes",
+        issues,
+      },
+      mockContext
+    );
+
+    // Now approve
+    await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test-project",
+        task_id: "bd-feedback-test",
+        worker_id: "worker-test",
+        status: "approved",
+        summary: "Fixed!",
+      },
+      mockContext
+    );
+
+    // Attempts should be cleared
+    const status = getReviewStatus("bd-feedback-test");
+    expect(status.attempt_count).toBe(0);
+    expect(status.remaining_attempts).toBe(3);
+  });
+
+  it("handles invalid issues JSON", async () => {
+    const result = await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test-project",
+        task_id: "bd-feedback-test",
+        worker_id: "worker-test",
+        status: "needs_changes",
+        issues: "not valid json",
+      },
+      mockContext
+    );
+
+    const parsed = JSON.parse(result);
+    expect(parsed.success).toBe(false);
+    expect(parsed.error).toContain("parse");
+  });
+
+  it("extracts epic ID from task ID for thread", async () => {
+    // Task ID format: bd-epic.subtask
+    const result = await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test-project",
+        task_id: "bd-epic-123.4",
+        worker_id: "worker-test",
+        status: "approved",
+      },
+      mockContext
+    );
+
+    const parsed = JSON.parse(result);
+    expect(parsed.success).toBe(true);
+    // The sendSwarmMessage mock was called with threadId = "bd-epic-123"
+  });
+});
+
+// ============================================================================
+// Integration: swarm_complete with review gate
+// ============================================================================
+
+describe("swarm_complete with review gate", () => {
+  // These tests verify the review gate behavior that was added to swarm_complete
+  // The actual swarm_complete tests are in swarm.integration.test.ts
+  // Here we test the review status functions that gate completion
+
+  beforeEach(() => {
+    clearReviewStatus("bd-gate-test");
+  });
+
+  it("isReviewApproved returns false for unreviewed task", () => {
+    expect(isReviewApproved("bd-gate-test")).toBe(false);
+  });
+
+  it("isReviewApproved returns true after markReviewApproved", () => {
+    markReviewApproved("bd-gate-test");
+    expect(isReviewApproved("bd-gate-test")).toBe(true);
+  });
+
+  it("getReviewStatus provides complete status info", () => {
+    const status = getReviewStatus("bd-gate-test");
+    expect(status).toEqual({
+      reviewed: false,
+      approved: false,
+      attempt_count: 0,
+      remaining_attempts: 3,
+    });
+  });
+
+  it("approval clears attempt count", async () => {
+    // Simulate some failed attempts
+    const issues = JSON.stringify([{ file: "x.ts", issue: "bug" }]);
+    await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test",
+        task_id: "bd-gate-test",
+        worker_id: "worker",
+        status: "needs_changes",
+        issues,
+      },
+      mockContext
+    );
+
+    let status = getReviewStatus("bd-gate-test");
+    expect(status.attempt_count).toBe(1);
+
+    // Approve
+    await swarm_review_feedback.execute(
+      {
+        project_key: "/tmp/test",
+        task_id: "bd-gate-test",
+        worker_id: "worker",
+        status: "approved",
+      },
+      mockContext
+    );
+
+    status = getReviewStatus("bd-gate-test");
+    expect(status.attempt_count).toBe(0);
+    expect(status.approved).toBe(true);
+  });
+});
+
+// ============================================================================
+// Worker prompt updates for review flow
+// ============================================================================
+
+describe("worker prompt with review instructions", () => {
+  // These tests verify that the review prompt includes proper instructions
+  // The actual worker prompt generation is in swarm-prompts.ts
+
+  it("review prompt includes response format for workers", () => {
+    const prompt = generateReviewPrompt({
+      epic_id: "bd-test",
+      epic_title: "Test Epic",
+      task_id: "bd-test.1",
+      task_title: "Test Task",
+      files_touched: [],
+      diff: "",
+    });
+
+    // Workers need to know how to respond
+    expect(prompt).toContain("Response Format");
+    expect(prompt).toContain("approved");
+    expect(prompt).toContain("needs_changes");
+  });
+
+  it("review prompt explains issue structure", () => {
+    const prompt = generateReviewPrompt({
+      epic_id: "bd-test",
+      epic_title: "Test Epic",
+      task_id: "bd-test.1",
+      task_title: "Test Task",
+      files_touched: [],
+      diff: "",
+    });
+
+    expect(prompt).toContain("file");
+    expect(prompt).toContain("line");
+    expect(prompt).toContain("issue");
+    expect(prompt).toContain("suggestion");
+  });
+});
+
+// ============================================================================
+// TDD ENFORCEMENT IN SWARM
+// ============================================================================
+
+describe("TDD enforcement in review criteria", () => {
+  it("review criteria includes test coverage check", () => {
+    const prompt = generateReviewPrompt({
+      epic_id: "bd-test",
+      epic_title: "Test Epic",
+      task_id: "bd-test.1",
+      task_title: "Test Task",
+      files_touched: ["src/foo.ts"],
+      diff: "+function foo() {}",
+    });
+
+    expect(prompt).toContain("Test Coverage");
+  });
+});
+
+// ============================================================================
+// Edge Cases
+// ============================================================================
+
+describe("edge cases", () => {
+  it("handles empty files_touched", () => {
+    const prompt = generateReviewPrompt({
+      epic_id: "bd-test",
+      epic_title: "Test Epic",
+      task_id: "bd-test.1",
+      task_title: "Test Task",
+      files_touched: [],
+      diff: "",
+    });
+
+    expect(prompt).toContain("## Files Modified");
+    // Should not crash, just have empty list
+  });
+
+  it("handles missing optional fields", () => {
+    const prompt = generateReviewPrompt({
+      epic_id: "bd-test",
+      epic_title: "Test Epic",
+      task_id: "bd-test.1",
+      task_title: "Test Task",
+      files_touched: [],
+      diff: "",
+      // No epic_description, task_description, dependencies, downstream
+    });
+
+    expect(prompt).toContain("Test Epic");
+    expect(prompt).toContain("Test Task");
+    // Should not include dependency sections
+    expect(prompt).not.toContain("This Task Builds On");
+    expect(prompt).not.toContain("Downstream Tasks");
+  });
+
+  it("handles special characters in diff", () => {
+    const prompt = generateReviewPrompt({
+      epic_id: "bd-test",
+      epic_title: "Test Epic",
+      task_id: "bd-test.1",
+      task_title: "Test Task",
+      files_touched: ["src/test.ts"],
+      diff: '+const regex = /[a-z]+/g;\n+const template = `Hello ${name}`;',
+    });
+
+    expect(prompt).toContain("regex");
+    expect(prompt).toContain("template");
+  });
+
+  it("handles very long diffs", () => {
+    const longDiff = "+line\n".repeat(1000);
+    const prompt = generateReviewPrompt({
+      epic_id: "bd-test",
+      epic_title: "Test Epic",
+      task_id: "bd-test.1",
+      task_title: "Test Task",
+      files_touched: ["src/big.ts"],
+      diff: longDiff,
+    });
+
+    // Should include the diff without truncation (truncation is caller's responsibility)
+    expect(prompt).toContain(longDiff);
+  });
+});