diff --git a/skills/creating-workflows/SKILL.md b/skills/creating-workflows/SKILL.md
index 058474b..824a4d9 100644
--- a/skills/creating-workflows/SKILL.md
+++ b/skills/creating-workflows/SKILL.md
@@ -37,6 +37,7 @@ steps:
 - **model** — Which model tier runs this step. Pick based on task complexity.
 - **prompt** — The instruction. Use `{{step-id}}` to reference earlier outputs. Use `{{input}}` for workflow input. Mutually exclusive with `command`.
 - **command** — Shell command run directly (no LLM). Output is captured and available via `{{step-id}}`. Costs $0. Mutually exclusive with `prompt`.
+- **expect** — Only for `command` steps. Values: `success` (step fails if exit code is non-zero) or `failure` (step fails if exit code is 0). Omit for the default behavior where all exit codes are informational. Enables bugfix reproduction gates: repro with `expect: failure` must fail before fix, verify with `expect: success` must pass after.
 - **parallel** — Lists step IDs to run concurrently. Results collected before next sequential step.
 - **loop** — Repeats with `max` iterations. Exits early if output contains `until` string. Optional `gate` command enforces quality after each iteration.
 - **loop.gate** — Shell command run after each loop iteration. Exit 0 = keep changes and commit. Non-zero = revert changes via `git checkout`. 3 consecutive gate failures trigger stuck detection and stop the loop.
diff --git a/src/engine/engine.go b/src/engine/engine.go
index b784ca5..56e6231 100644
--- a/src/engine/engine.go
+++ b/src/engine/engine.go
@@ -297,6 +297,24 @@ func (e *Engine) runStep(ctx context.Context, step *WfStep, session *lib.Session
 			return 0, "", fmt.Errorf("step %s command failed: %w", step.ID, err)
 		}
 
+		// Check expect condition on exit code.
+		if step.Expect == "failure" && exitCode == 0 {
+			reason := "expected failure but got exit code 0"
+			dbStep.Status = "failed"
+			dbStep.ChangeSummary = reason
+			e.db.UpdateStep(dbStep)
+			fmt.Printf("  %s\n\n", reason)
+			return 0, "", fmt.Errorf("step %s: %s", step.ID, reason)
+		}
+		if step.Expect == "success" && exitCode != 0 {
+			reason := fmt.Sprintf("expected success but got exit code %d", exitCode)
+			dbStep.Status = "failed"
+			dbStep.ChangeSummary = reason
+			e.db.UpdateStep(dbStep)
+			fmt.Printf("  %s\n\n", reason)
+			return 0, "", fmt.Errorf("step %s: %s", step.ID, reason)
+		}
+
 		// Include exit code in output so downstream steps can check it
 		fullOutput := fmt.Sprintf("%s\nexit code: %d", strings.TrimRight(output, "\n"), exitCode)
 
diff --git a/src/engine/engine_test.go b/src/engine/engine_test.go
index fcdccc1..e89446f 100644
--- a/src/engine/engine_test.go
+++ b/src/engine/engine_test.go
@@ -927,6 +927,193 @@ func TestRunWorkflowCommandNonZeroExit(t *testing.T) {
 	}
 }
 
+// ---------------------------------------------------------------------------
+// Expect field tests
+// ---------------------------------------------------------------------------
+
+func TestRunWorkflowExpectFailure(t *testing.T) {
+	db := tempDB(t)
+	dir, git := initGitRepo(t)
+	runner := newMockRunner(nil, nil)
+	eng := mustEngine(t, db, git, runner, dir)
+
+	// expect: failure — non-zero exit should succeed (bug repro)
+	wf := &Workflow{
+		Name: "test",
+		Steps: []WfStep{
+			{ID: "repro", Command: "exit 1", Expect: "failure"},
+		},
+	}
+
+	res, err := eng.RunWorkflow(context.Background(), wf, RunConfig{Input: "test"})
+	if err != nil {
+		t.Fatalf("expected success for non-zero exit with expect:failure, got: %v", err)
+	}
+	output, ok := res.Outputs["repro"]
+	if !ok {
+		t.Fatal("repro output missing")
+	}
+	if !strings.Contains(output, "exit code: 1") {
+		t.Errorf("output should contain exit code, got %q", output)
+	}
+}
+
+func TestRunWorkflowExpectFailureButGotZero(t *testing.T) {
+	db := tempDB(t)
+	dir, git := initGitRepo(t)
+	runner := newMockRunner(nil, nil)
+	eng := mustEngine(t, db, git, runner, dir)
+
+	// expect: failure — zero exit should FAIL (bug not reproducible)
+	wf := &Workflow{
+		Name: "test",
+		Steps: []WfStep{
+			{ID: "repro", Command: "exit 0", Expect: "failure"},
+		},
+	}
+
+	_, err := eng.RunWorkflow(context.Background(), wf, RunConfig{Input: "test"})
+	if err == nil {
+		t.Fatal("expected error for exit 0 with expect:failure, got nil")
+	}
+	if !strings.Contains(err.Error(), "expected failure") {
+		t.Errorf("error = %q, want it to contain 'expected failure'", err.Error())
+	}
+}
+
+func TestRunWorkflowExpectSuccessPass(t *testing.T) {
+	db := tempDB(t)
+	dir, git := initGitRepo(t)
+	runner := newMockRunner(nil, nil)
+	eng := mustEngine(t, db, git, runner, dir)
+
+	// expect: success — exit 0 should succeed
+	wf := &Workflow{
+		Name: "test",
+		Steps: []WfStep{
+			{ID: "check", Command: "exit 0", Expect: "success"},
+		},
+	}
+
+	res, err := eng.RunWorkflow(context.Background(), wf, RunConfig{Input: "test"})
+	if err != nil {
+		t.Fatalf("expected success for exit 0 with expect:success, got: %v", err)
+	}
+	if !strings.Contains(res.Outputs["check"], "exit code: 0") {
+		t.Error("exit code not in output")
+	}
+}
+
+func TestRunWorkflowExpectSuccessButGotNonZero(t *testing.T) {
+	db := tempDB(t)
+	dir, git := initGitRepo(t)
+	runner := newMockRunner(nil, nil)
+	eng := mustEngine(t, db, git, runner, dir)
+
+	// expect: success — non-zero exit should FAIL
+	wf := &Workflow{
+		Name: "test",
+		Steps: []WfStep{
+			{ID: "check", Command: "exit 1", Expect: "success"},
+		},
+	}
+
+	_, err := eng.RunWorkflow(context.Background(), wf, RunConfig{Input: "test"})
+	if err == nil {
+		t.Fatal("expected error for exit 1 with expect:success, got nil")
+	}
+	if !strings.Contains(err.Error(), "expected success") {
+		t.Errorf("error = %q, want it to contain 'expected success'", err.Error())
+	}
+}
+
+func TestRunWorkflowNoExpectDefault(t *testing.T) {
+	db := tempDB(t)
+	dir, git := initGitRepo(t)
+	runner := newMockRunner(nil, nil)
+	eng := mustEngine(t, db, git, runner, dir)
+
+	// No expect field (default) — non-zero exit is informational, not fatal
+	wf := &Workflow{
+		Name: "test",
+		Steps: []WfStep{
+			{ID: "check", Command: "exit 1"},
+		},
+	}
+
+	res, err := eng.RunWorkflow(context.Background(), wf, RunConfig{Input: "test"})
+	if err != nil {
+		t.Fatalf("default (no expect) should not fail on non-zero exit: %v", err)
+	}
+	if !strings.Contains(res.Outputs["check"], "exit code: 1") {
+		t.Error("exit code not in output")
+	}
+}
+
+func TestParseExpectField(t *testing.T) {
+	yaml := `
+name: Expect Test
+description: test
+
+steps:
+  - id: repro
+    command: "npm test"
+    expect: failure
+  - id: verify
+    command: "npm test"
+    expect: success
+`
+	wf, err := Parse([]byte(yaml))
+	if err != nil {
+		t.Fatalf("Parse: %v", err)
+	}
+	if wf.Steps[0].Expect != "failure" {
+		t.Errorf("step 0 expect = %q, want failure", wf.Steps[0].Expect)
+	}
+	if wf.Steps[1].Expect != "success" {
+		t.Errorf("step 1 expect = %q, want success", wf.Steps[1].Expect)
+	}
+}
+
+func TestValidateExpectOnPromptStep(t *testing.T) {
+	yaml := `
+name: Bad
+description: test
+
+steps:
+  - id: broken
+    model: smart
+    prompt: "do something"
+    expect: failure
+`
+	_, err := Parse([]byte(yaml))
+	if err == nil {
+		t.Fatal("expected validation error for expect on prompt step")
+	}
+	if !strings.Contains(err.Error(), "expect without command") {
+		t.Errorf("error = %q, want 'expect without command'", err.Error())
+	}
+}
+
+func TestValidateExpectInvalidValue(t *testing.T) {
+	yaml := `
+name: Bad
+description: test
+
+steps:
+  - id: broken
+    command: "echo hi"
+    expect: maybe
+`
+	_, err := Parse([]byte(yaml))
+	if err == nil {
+		t.Fatal("expected validation error for invalid expect value")
+	}
+	if !strings.Contains(err.Error(), "invalid expect") {
+		t.Errorf("error = %q, want 'invalid expect'", err.Error())
+	}
+}
+
 // ---------------------------------------------------------------------------
 // Gate tests
 // ---------------------------------------------------------------------------
diff --git a/src/engine/workflow.go b/src/engine/workflow.go
index af32d6e..0eff7e0 100644
--- a/src/engine/workflow.go
+++ b/src/engine/workflow.go
@@ -31,6 +31,7 @@ type WfStep struct {
 	Model    string   `yaml:"model"`
 	Prompt   string   `yaml:"prompt"`
 	Command  string   `yaml:"command"`
+	Expect   string   `yaml:"expect"`
 	Parallel []string `yaml:"parallel"`
 	Loop     *Loop    `yaml:"loop"`
 	Branch   []Branch `yaml:"branch"`
@@ -101,6 +102,12 @@ func validate(wf *Workflow) error {
 		if len(s.Parallel) > 0 && (s.Prompt != "" || s.Command != "") {
 			return fmt.Errorf("step %q has both parallel and prompt/command — these are mutually exclusive", s.ID)
 		}
+		if s.Expect != "" && s.Command == "" {
+			return fmt.Errorf("step %q has expect without command — expect only applies to command steps", s.ID)
+		}
+		if s.Expect != "" && s.Expect != "success" && s.Expect != "failure" {
+			return fmt.Errorf("step %q has invalid expect %q — must be \"success\" or \"failure\"", s.ID, s.Expect)
+		}
 		if s.Command != "" && s.Loop != nil {
 			return fmt.Errorf("step %q has both command and loop — these are mutually exclusive", s.ID)
 		}