prizmkit 1.1.114 → 1.1.115

package/bundled/dev-pipeline/tests/test_generate_bootstrap_prompt.py

@@ -215,7 +215,6 @@ class TestExplicitProjectRoot:
             output=str(main_checkout / ".prizmkit" / "state" / "features" / "F-001" / "sessions" / "session-1" / "bootstrap-prompt.md"),
             template=None,
             mode=None,
-            critic=None,
             extract_baselines=False,
             project_root=str(execution_root),
         )
@@ -403,23 +402,6 @@ class TestProcessModeBlocks:
         result = process_mode_blocks(tpl, "standard", init_done=False)
         assert "need init" in result
 
-    def test_critic_enabled_keeps_critic_block(self):
-        tpl = "before\n{{IF_CRITIC_ENABLED}}\ncritic content\n{{END_IF_CRITIC_ENABLED}}\nafter"
-        result = process_mode_blocks(tpl, "standard", init_done=True, critic_enabled=True)
-        assert "critic content" in result
-        assert "IF_CRITIC_ENABLED" not in result
-
-    def test_critic_disabled_removes_critic_block(self):
-        tpl = "before\n{{IF_CRITIC_ENABLED}}\ncritic content\n{{END_IF_CRITIC_ENABLED}}\nafter"
-        result = process_mode_blocks(tpl, "standard", init_done=True, critic_enabled=False)
-        assert "critic content" not in result
-        assert "before" in result
-        assert "after" in result
-
-    def test_critic_default_is_disabled(self):
-        tpl = "{{IF_CRITIC_ENABLED}}critic{{END_IF_CRITIC_ENABLED}}"
-        result = process_mode_blocks(tpl, "standard", init_done=True)
-        assert "critic" not in result
 
 
 
@@ -435,7 +417,6 @@ class TestScopedFeatureTestGate:
             str(sections_dir),
             init_done=True,
             is_resume=False,
-            critic_enabled=False,
             browser_enabled=False,
         )
         names = [name for name, _ in sections]
@@ -470,7 +451,6 @@ class TestScopedFeatureTestGate:
             str(sections_dir),
             init_done=True,
             is_resume=False,
-            critic_enabled=False,
             browser_enabled=False,
         )
         rendered = "\n".join(content for _, content in sections)
@@ -490,7 +470,6 @@ class TestScopedFeatureTestGate:
             str(sections_dir),
             init_done=True,
             is_resume=False,
-            critic_enabled=False,
             browser_enabled=False,
         )
         names = [name for name, _ in sections]
@@ -506,7 +485,6 @@ class TestScopedFeatureTestGate:
             str(sections_dir),
             init_done=True,
             is_resume=False,
-            critic_enabled=False,
             browser_enabled=False,
         )
         checkpoint = generate_checkpoint_definition(
@@ -853,7 +831,7 @@ class TestHeadlessRecoveryReframing:
             rendered = "\n".join(
                 section for _name, section in assemble_sections(
                     mode, str(sections_dir), init_done=True, is_resume=False,
-                    critic_enabled=False, browser_enabled=False,
+                    browser_enabled=False,
                 )
             )
             assert deleted_title not in rendered
@@ -883,7 +861,7 @@ class TestDirectOrchestratorImplementationPrompts:
         assert "{{AGENT_PROMPT_DEV_IMPLEMENT}}" not in replacements
         assert FORBIDDEN_DEV_FIX_PLACEHOLDER not in replacements
         assert FORBIDDEN_DEV_RESUME_PLACEHOLDER not in replacements
-        assert "{{AGENT_PROMPT_CRITIC_PLAN_CHALLENGE}}" in replacements
+        assert ACTIVE_AGENT_PROMPTS == {}
         assert "{{AGENT_PROMPT_REVIEWER_REVIEW}}" not in replacements
         assert (templates_dir / "agent-prompts" / "reviewer-review.md").exists() is False
         assert (templates_dir / "agent-prompts" / "dev-implement.md").exists() is False
@@ -897,7 +875,7 @@ class TestDirectOrchestratorImplementationPrompts:
             rendered = "\n".join(
                 section for _name, section in assemble_sections(
                     mode, str(sections_dir), init_done=True, is_resume=False,
-                    critic_enabled=(mode == "full"), browser_enabled=False,
+                    browser_enabled=False,
                 )
             )
 
@@ -963,9 +941,9 @@ class TestContinuationHandoff:
             continuation_summary_path="",
         )
         checkpoint = {"steps": [
-            {"id": "S01", "skill": "test-baseline", "status": "completed"},
-            {"id": "S02", "skill": "context-snapshot", "status": "skipped"},
-            {"id": "S03", "skill": "prizmkit-implement", "status": "pending"},
+            {"id": "S01", "skill": "test-baseline", "name": "Test Baseline", "status": "completed"},
+            {"id": "S02", "skill": "context-snapshot", "name": "Context Snapshot", "status": "skipped"},
+            {"id": "S03", "skill": "prizmkit-implement", "name": "Implement", "status": "pending"},
         ]}
 
         rendered = append_continuation_handoff(
@@ -1000,15 +978,15 @@ class TestContinuationHandoff:
 
     def test_checkpoint_cursor_prefers_in_progress_before_pending(self):
         checkpoint = {"steps": [
-            {"id": "S01", "skill": "done", "status": "completed"},
-            {"id": "S02", "skill": "active", "status": "in_progress"},
-            {"id": "S03", "skill": "later", "status": "pending"},
+            {"id": "S01", "skill": "done", "name": "Done", "status": "completed"},
+            {"id": "S02", "skill": "active", "name": "Active", "status": "in_progress", "depends_on": "S01"},
+            {"id": "S03", "skill": "later", "name": "Later", "status": "pending", "depends_on": "S02"},
         ]}
         assert checkpoint_cursor(checkpoint)["first_cursor"] == "S02 (active)"
 
 
 # ---------------------------------------------------------------------------
-# F-033 headless prompt cleanup and platform-aware critic paths
+# F-033 headless prompt cleanup and platform-aware paths
 # ---------------------------------------------------------------------------
 
 HEADLESS_FORBIDDEN_PROMPT_STRINGS = [
@@ -1030,7 +1008,7 @@ def _write_feature_list(path, feature):
     path.write_text(json.dumps({"features": [feature], "global_context": {"language": "Python"}}), encoding="utf-8")
 
 
-def _render_feature_prompt(tmp_path, mode="standard", critic=False, template=None, platform="claude", browser_interaction=None):
+def _render_feature_prompt(tmp_path, mode="standard", template=None, platform="claude", browser_interaction=None):
     project = tmp_path / "project"
     project.mkdir(parents=True)
     (project / ".prizmkit" / "plans").mkdir(parents=True)
@@ -1040,7 +1018,7 @@ def _render_feature_prompt(tmp_path, mode="standard", critic=False, template=Non
     for platform_dir, suffix in ((".claude", ".md"), (".codex", ".toml")):
         agents_dir = project / platform_dir / "agents"
         agents_dir.mkdir(parents=True)
-        for name in ("dev", "reviewer", "critic"):
+        for name in ("dev", "reviewer"):
             (agents_dir / f"prizm-dev-team-{name}{suffix}").write_text("agent\n", encoding="utf-8")
     feature = {
         "id": "F-123",
@@ -1048,7 +1026,6 @@ def _render_feature_prompt(tmp_path, mode="standard", critic=False, template=Non
         "description": "Clean prompt text",
         "estimated_complexity": {"lite": "low", "standard": "medium", "full": "high"}[mode],
         "acceptance_criteria": ["Prompts are clean"],
-        "critic": critic,
     }
     if browser_interaction is not None:
         feature["browser_interaction"] = browser_interaction
@@ -1068,8 +1045,6 @@ def _render_feature_prompt(tmp_path, mode="standard", critic=False, template=Non
         "--output", str(output),
         "--mode", mode,
     ]
-    if critic:
-        args.extend(["--critic", "true"])
     if template:
         args.extend(["--template", str(Path("dev-pipeline/templates") / template)])
     old_argv = os.sys.argv
@@ -1096,7 +1071,7 @@ def _ensure_agent_files(project, platforms=("claude", "codex")):
         suffix = ".toml" if platform == "codex" else ".md"
         agents_dir = project / platform_dir / "agents"
         agents_dir.mkdir(parents=True, exist_ok=True)
-        for name in ("dev", "reviewer", "critic"):
+        for name in ("dev", "reviewer"):
             (agents_dir / f"prizm-dev-team-{name}{suffix}").write_text("agent\n", encoding="utf-8")
 
 
@@ -1280,7 +1255,7 @@ class TestHeadlessPromptCleanupF033:
     def test_rendered_feature_bugfix_refactor_prompts_are_clean(self, tmp_path):
         prompts = []
         for mode in ("lite", "standard", "full"):
-            prompts.append(_render_feature_prompt(tmp_path / mode, mode=mode, critic=(mode != "lite")))
+            prompts.append(_render_feature_prompt(tmp_path / mode, mode=mode))
         prompts.append(_render_bugfix_prompt(tmp_path))
         prompts.append(_render_refactor_prompt(tmp_path))
 
@@ -1309,7 +1284,7 @@ class TestHeadlessPromptCleanupF033:
 
     def test_legacy_feature_templates_are_clean_when_rendered(self, tmp_path):
         for template, mode in (("bootstrap-tier1.md", "lite"), ("bootstrap-tier2.md", "standard"), ("bootstrap-tier3.md", "full")):
-            prompt = _render_feature_prompt(tmp_path / template, mode=mode, critic=(mode != "lite"), template=template)
+            prompt = _render_feature_prompt(tmp_path / template, mode=mode, template=template)
             for forbidden in HEADLESS_FORBIDDEN_PROMPT_STRINGS:
                 assert forbidden not in prompt
             assert "prizmkit-test-gate.py" not in prompt
@@ -1342,29 +1317,13 @@ class TestHeadlessPromptCleanupF033:
             prompt = _render_feature_prompt(
                 tmp_path / f"browser-{template}",
                 mode=mode,
-                critic=(mode != "lite"),
-                template=template,
+                                template=template,
                 browser_interaction={"tool": "auto", "verify_steps": ["Open the page"]},
             )
             assert "prizmkit-runtime-helper.py" in prompt
             for snippet in forbidden:
                 assert snippet not in prompt
 
-    def test_claude_config_wins_over_codex_directory_for_critic_path(self, tmp_path, monkeypatch):
-        project = tmp_path / "mixed"
-        (project / ".claude" / "agents").mkdir(parents=True)
-        (project / ".codex" / "agents").mkdir(parents=True)
-        (project / ".prizmkit").mkdir()
-        (project / ".prizmkit" / "config.json").write_text(
-            json.dumps({"platform": "claude", "ai_cli": "claude"}), encoding="utf-8"
-        )
-        monkeypatch.delenv("PRIZMKIT_PLATFORM", raising=False)
-
-        assert resolve_prompt_platform(str(project)) == "claude"
-        prompt = _render_feature_prompt(tmp_path / "rendered-mixed", mode="standard", critic=True, platform="claude")
-        assert ".claude/agents/prizm-dev-team-critic.md" in prompt
-        assert ".codex/agents/prizm-dev-team-critic.toml" not in prompt
-
     def test_bugfix_refactor_share_claude_platform_resolution(self, tmp_path, monkeypatch):
         from generate_bugfix_prompt import build_replacements as bugfix_build_replacements
 
@@ -1380,7 +1339,7 @@ class TestHeadlessPromptCleanupF033:
             bug_id="B-001", session_id="session-1", run_id="run-1",
             retry_count="0", resume_phase="null", state_dir=None,
             output="out.md", project_root=str(bug_project), template=None,
-            mode=None, critic=None,
+            mode=None,
         )
         replacements = bugfix_build_replacements(
             args, {"id": "B-001", "title": "Broken"}, {}, str(Path("dev-pipeline/scripts").resolve())
@@ -1401,7 +1360,7 @@ class TestHeadlessPromptCleanupF033:
             bug_id="B-001", session_id="session-1", run_id="run-1",
             retry_count="0", resume_phase="null", state_dir=None,
             output="out.md", project_root=str(bug_project), template=None,
-            mode=None, critic=None,
+            mode=None,
         )
         replacements = bugfix_build_replacements(
             args, {"id": "B-001", "title": "Broken"}, {}, str(Path("dev-pipeline/scripts").resolve())
@@ -1411,15 +1370,6 @@ class TestHeadlessPromptCleanupF033:
         assert replacements["{{REVIEWER_SUBAGENT_PATH}}"].endswith(".codex/agents/prizm-dev-team-reviewer.toml")
         assert ".claude/agents/prizm-dev-team-reviewer.md" not in refactor_prompt
 
-    def test_explicit_codex_env_still_renders_codex_critic_path(self, tmp_path, monkeypatch):
-        project = tmp_path / "codex"
-        (project / ".claude" / "agents").mkdir(parents=True)
-        (project / ".codex" / "agents").mkdir(parents=True)
-        monkeypatch.setenv("PRIZMKIT_PLATFORM", "codex")
-
-        assert resolve_prompt_platform(str(project)) == "codex"
-
-
 
 class TestFeatureBootstrapShellExtraction:
     RETIRED_EXTRACTED_SNIPPETS = [
@@ -1465,8 +1415,7 @@ class TestFeatureBootstrapShellExtraction:
             prompt = _render_feature_prompt(
                 tmp_path / f"section-{mode}",
                 mode=mode,
-                critic=(mode != "lite"),
-                browser_interaction={"tool": "auto", "verify_steps": ["Open the page"]},
+                                browser_interaction={"tool": "auto", "verify_steps": ["Open the page"]},
             )
 
             self.assert_shell_extracted_operations_removed(prompt)
@@ -1488,8 +1437,7 @@ class TestFeatureBootstrapShellExtraction:
             prompt = _render_feature_prompt(
                 tmp_path / f"legacy-{template}",
                 mode=mode,
-                critic=(mode != "lite"),
-                template=template,
+                                template=template,
                 browser_interaction={"tool": "playwright-cli", "verify_steps": ["Open the page"]},
             )
 
@@ -1512,8 +1460,7 @@ class TestFeatureBootstrapShellExtraction:
         tier2 = _render_feature_prompt(
             tmp_path / "legacy-tier2-helper-map",
             mode="standard",
-            critic=True,
-            template="bootstrap-tier2.md",
+                        template="bootstrap-tier2.md",
         )
         assert "artifact exists .prizmkit/specs/123-prompt-cleanup/failure-log.md" in tier2
         assert "artifact tree . --max-depth 2" in tier2
@@ -1522,8 +1469,7 @@ class TestFeatureBootstrapShellExtraction:
         tier3 = _render_feature_prompt(
             tmp_path / "legacy-tier3-helper-map",
             mode="full",
-            critic=True,
-            template="bootstrap-tier3.md",
+                        template="bootstrap-tier3.md",
         )
         assert "text search . --pattern \"123-prompt-cleanup\" --json" in tier3
         assert "text count .prizmkit/specs/123-prompt-cleanup/plan.md --pattern \"- [ ]\"" in tier3
@@ -1619,3 +1565,29 @@ class TestDetectTestCommands:
         # Should detect test:unit as a viable test script
         assert "npm run test:unit" in result
 
+
+
+def test_feature_checkpoint_generation_skips_noop_and_expands_only_finalization_group():
+    from generate_bootstrap_prompt import generate_checkpoint_definition
+
+    checkpoint = generate_checkpoint_definition(
+        [
+            ("phase0-test-baseline", "noop"),
+            ("phase-context-snapshot", "context"),
+            ("phase-commit", "finalize"),
+        ],
+        "standard",
+        "feature-pipeline",
+        "F-123",
+        "123-prompt-cleanup",
+        "session-1",
+    )
+
+    assert [step["skill"] for step in checkpoint["steps"]] == [
+        "context-snapshot",
+        "prizmkit-retrospective",
+        "prizmkit-committer",
+        "completion-summary",
+    ]
+    assert checkpoint["steps"][0]["depends_on"] is None
+    assert checkpoint["steps"][1]["depends_on"] == "S01"

package/bundled/dev-pipeline/tests/test_generate_bugfix_prompt.py

@@ -12,6 +12,7 @@ from generate_bugfix_prompt import (
     format_global_context,
     format_error_source_details,
     format_environment,
+    generate_bugfix_checkpoint,
     process_conditional_blocks,
     render_template,
 )
@@ -178,7 +179,6 @@ class TestExplicitProjectRoot:
             output=str(main_checkout / ".prizmkit" / "state" / "bugfix" / "B-001" / "sessions" / "session-1" / "bootstrap-prompt.md"),
             template=None,
             mode=None,
-            critic=None,
             project_root=str(execution_root),
         )
         bug = {"id": "B-001", "title": "Worktree Bug", "description": "d"}
@@ -275,7 +275,7 @@ class TestBugfixContinuationHandoff:
         )
         prompt = append_continuation_handoff(
             "BUGFIX PROMPT", args, str(tmp_path), "bugfix", "B-007", "B-007",
-            {"steps": [{"id": "S01", "skill": "diagnose", "status": "pending"}]},
+            {"steps": [{"id": "S01", "skill": "diagnose", "name": "Diagnose", "status": "pending"}]},
         )
 
         assert prompt.startswith("BUGFIX PROMPT")
@@ -408,7 +408,7 @@ def _ensure_agent_files(project, platforms=None):
         suffix = ".toml" if plat == "codex" else ".md"
         agents_dir = project / platform_dir / "agents"
         agents_dir.mkdir(parents=True, exist_ok=True)
-        for name in ("dev", "reviewer", "critic"):
+        for name in ("dev", "reviewer"):
             (agents_dir / f"prizm-dev-team-{name}{suffix}").write_text("agent\n", encoding="utf-8")
 
 
@@ -439,3 +439,25 @@ class TestBugfixBootstrapShellExtraction:
         content = Path("dev-pipeline/scripts/generate-bugfix-prompt.py").read_text(encoding="utf-8")
         for snippet in RETIRED_EXTRACTED_SNIPPETS_BUGFIX:
             assert snippet not in content, f"Shell snippet in generate-bugfix-prompt.py: {snippet}"
+
+
+class TestBugfixCheckpointGeneration:
+    def test_setup_section_skips_and_finalization_group_expands(self):
+        checkpoint = generate_bugfix_checkpoint(
+            "B-001",
+            "session-1",
+            sections=[
+                ("bugfix-phase-init", "setup"),
+                ("bugfix-phase-implement", "implement"),
+                ("bugfix-phase-commit-report", "finalize"),
+            ],
+        )
+
+        assert [step["skill"] for step in checkpoint["steps"]] == [
+            "prizmkit-init",
+            "prizmkit-implement",
+            "prizmkit-retrospective",
+            "prizmkit-committer",
+            "bug-report",
+        ]
+        assert checkpoint["steps"][2]["depends_on"] == "S02"

package/bundled/dev-pipeline/tests/test_generate_refactor_prompt.py

@@ -30,7 +30,7 @@ class TestRefactorContinuationHandoff:
         )
         prompt = append_continuation_handoff(
             "REFACTOR PROMPT", args, str(tmp_path), "refactor", "R-007", "R-007",
-            {"steps": [{"id": "S01", "skill": "plan", "status": "pending"}]},
+            {"steps": [{"id": "S01", "skill": "plan", "name": "Plan", "status": "pending"}]},
         )
 
         assert prompt.startswith("REFACTOR PROMPT")
@@ -74,7 +74,6 @@ class TestExplicitProjectRoot:
             output=str(project / "out.md"),
             template=None,
             mode=None,
-            critic=None,
         )
         refactor = {"id": "R-001", "title": "Restructure", "description": "d"}
 
@@ -109,7 +108,6 @@ class TestExplicitProjectRoot:
             output=str(main_checkout / ".prizmkit" / "state" / "refactor" / "R-001" / "sessions" / "session-1" / "bootstrap-prompt.md"),
             template=None,
             mode=None,
-            critic=None,
             project_root=str(execution_root),
         )
         refactor = {"id": "R-001", "title": "Worktree Refactor", "description": "d"}
@@ -161,7 +159,6 @@ class TestRefactorPromptMaxLogRemoval:
             output="out.md",
             template=None,
             mode=None,
-            critic=None,
         )
         refactor = {"id": "R-001", "title": "Restructure", "description": "d"}
 
@@ -243,7 +240,7 @@ def _render_refactor_prompt_full(tmp_path, platform="claude"):
         suffix = ".toml" if plat == "codex" else ".md"
         agents_dir = project / platform_dir / "agents"
         agents_dir.mkdir(parents=True, exist_ok=True)
-        for name in ("dev", "reviewer", "critic"):
+        for name in ("dev", "reviewer"):
             (agents_dir / f"prizm-dev-team-{name}{suffix}").write_text("agent\n", encoding="utf-8")
 
     refactor_list = project / ".prizmkit" / "plans" / "refactor-list.json"
@@ -318,3 +315,26 @@ class TestRefactorBootstrapShellExtraction:
         content = Path("dev-pipeline/scripts/generate-refactor-prompt.py").read_text(encoding="utf-8")
         for snippet in RETIRED_EXTRACTED_SNIPPETS_REFACTOR:
             assert snippet not in content, f"Shell snippet in generate-refactor-prompt.py: {snippet}"
+
+
+class TestRefactorCheckpointGeneration:
+    def test_setup_section_is_step_and_finalization_group_expands(self):
+        checkpoint = generate_refactor_checkpoint(
+            "R-001",
+            "session-1",
+            sections=[
+                ("refactor-phase-init", "setup"),
+                ("refactor-phase-implement", "implement"),
+                ("refactor-phase-commit-report", "finalize"),
+            ],
+        )
+
+        assert [step["skill"] for step in checkpoint["steps"]] == [
+            "prizmkit-init",
+            "prizmkit-implement",
+            "prizmkit-retrospective",
+            "prizmkit-committer",
+            "refactor-report",
+            "completion-summary",
+        ]
+        assert checkpoint["steps"][2]["depends_on"] == "S02"

package/bundled/dev-pipeline/tests/test_python_runner_parity.py

@@ -320,7 +320,7 @@ def test_runner_models_parse_legacy_args_and_environment(tmp_path):
     invocation = parse_invocation(
         family,
         "run",
-        ("custom-list.json", "F-2", "--features", "F-003:F-001", "--max-retries", "7", "--max-infra-retries=5", "--mode", "full", "--critic"),
+        ("custom-list.json", "F-2", "--features", "F-003:F-001", "--max-retries", "7", "--max-infra-retries=5", "--mode", "full"),
     )
     env = RunnerEnvironment.from_env(
         {
@@ -340,7 +340,6 @@ def test_runner_models_parse_legacy_args_and_environment(tmp_path):
     assert invocation.max_retries == 7
     assert invocation.max_infra_retries == 5
     assert invocation.mode == "full"
-    assert invocation.critic is True
     assert env.stop_on_failure is True
     assert env.auto_push is True
     assert env.dev_branch == "dev/custom"
@@ -765,7 +764,6 @@ def test_prompt_generation_result_preserves_per_item_model(monkeypatch, tmp_path
                     "model": "per-item-model",
                     "pipeline_mode": "standard",
                     "agent_count": 3,
-                    "critic_enabled": False,
                 }
             ),
             stderr="",
@@ -815,7 +813,6 @@ def test_prompt_generation_passes_explicit_execution_root_to_generator(monkeypat
                     "model": "",
                     "pipeline_mode": "lite",
                     "agent_count": 1,
-                    "critic_enabled": False,
                 }
             ),
             stderr="",
@@ -881,7 +878,6 @@ def _install_runner_session_fakes(monkeypatch, runners, *, statuses=("success",)
             model="per-item-model",
             pipeline_mode="lite",
             agent_count=1,
-            critic_enabled=False,
             checkpoint_path=execution_root / ".prizmkit" / "specs" / "001-low" / "workflow-checkpoint.json",
             artifact_path=execution_root / ".prizmkit" / "specs" / "001-low",
             raw={},
@@ -1144,7 +1140,6 @@ def test_explicit_worktree_materializes_ignored_support_assets_before_prompt_and
             model="",
             pipeline_mode="lite",
             agent_count=1,
-            critic_enabled=False,
             checkpoint_path=execution_root / ".prizmkit" / "specs" / "001-support" / "workflow-checkpoint.json",
             artifact_path=execution_root / ".prizmkit" / "specs" / "001-support",
             raw={},
@@ -1294,7 +1289,7 @@ def test_no_worktree_recovers_completed_pending_branch_before_new_session(monkey
     artifact_dir = paths.project_root / ".prizmkit" / "specs" / "001-low"
     artifact_dir.mkdir(parents=True)
     (artifact_dir / "workflow-checkpoint.json").write_text(
-        json.dumps({"steps": [{"id": "S01", "skill": "done", "status": "completed"}]}),
+        json.dumps({"steps": [{"id": "S01", "skill": "done", "name": "Done", "status": "completed"}]}),
         encoding="utf-8",
     )
     updates = []
@@ -1530,7 +1525,7 @@ def test_foreground_runner_emits_structured_feature_progress(monkeypatch, tmp_pa
     assert "Code retry: 1 / 3" in captured.err
     assert "Preflight: no user-visible changes to preserve before feature task F-001" in captured.err
     assert "Pipeline mode: lite (Tier 1 — Single Agent)" in captured.err
-    assert "Agents: 1 (critic: disabled)" in captured.err
+    assert "Agents: 1" in captured.err
     assert "Spawning AI CLI session: F-001-session-1" in captured.err
     assert "Session result: success" in captured.err
 
@@ -1667,7 +1662,7 @@ def test_classification_accepts_late_infra_error_with_completed_checkpoint_and_c
     artifact_dir = tmp_path / ".prizmkit" / "specs" / "001-done"
     artifact_dir.mkdir(parents=True)
     checkpoint = artifact_dir / "workflow-checkpoint.json"
-    checkpoint.write_text(json.dumps({"steps": [{"id": "S01", "skill": "done", "status": "completed"}]}), encoding="utf-8")
+    checkpoint.write_text(json.dumps({"steps": [{"id": "S01", "skill": "done", "name": "Done", "status": "completed"}]}), encoding="utf-8")
     log = tmp_path / "session.log"
     log.write_text("model quota exceeded; API error after commit", encoding="utf-8")
     raw = AISessionResult(
@@ -1792,7 +1787,7 @@ def test_classification_accepts_terminal_success_with_completed_checkpoint_witho
     artifact_dir = tmp_path / ".prizmkit" / "specs" / "001-done"
     artifact_dir.mkdir(parents=True)
     checkpoint = artifact_dir / "workflow-checkpoint.json"
-    checkpoint.write_text(json.dumps({"steps": [{"id": "S01", "skill": "done", "status": "completed"}]}), encoding="utf-8")
+    checkpoint.write_text(json.dumps({"steps": [{"id": "S01", "skill": "done", "name": "Done", "status": "completed"}]}), encoding="utf-8")
     (artifact_dir / "completion-summary.json").write_text("{}", encoding="utf-8")
     log = tmp_path / "session.log"
     log.write_text("terminal success", encoding="utf-8")
@@ -2207,3 +2202,126 @@ def test_python_daemon_progress_uses_list_status_without_runtime_files(tmp_path)
         else:
             assert summary["auto_skipped"] == 1
 
+
+
+def _init_git_repo_for_invalid_checkpoint_test(tmp_path):
+    subprocess.run(["git", "init", "-b", "main"], cwd=tmp_path, check=True, stdout=subprocess.DEVNULL)
+    subprocess.run(["git", "config", "user.email", "test@example.com"], cwd=tmp_path, check=True)
+    subprocess.run(["git", "config", "user.name", "Test"], cwd=tmp_path, check=True)
+    (tmp_path / "base.txt").write_text("base", encoding="utf-8")
+    subprocess.run(["git", "add", "base.txt"], cwd=tmp_path, check=True)
+    subprocess.run(["git", "commit", "-m", "base"], cwd=tmp_path, check=True, stdout=subprocess.DEVNULL)
+    return subprocess.run(["git", "rev-parse", "HEAD"], cwd=tmp_path, check=True, stdout=subprocess.PIPE, text=True).stdout.strip()
+
+
+def _terminal_success_result(tmp_path):
+    from prizmkit_runtime.sessions import AISessionCommand, AISessionResult
+    from prizmkit_runtime.status import ProgressSummary
+
+    log = tmp_path / "session.log"
+    log.write_text("terminal success", encoding="utf-8")
+    return AISessionResult(
+        command=AISessionCommand(("fake",), None, "claude", "stdin"),
+        exit_code=0,
+        timed_out=False,
+        termination_reason=None,
+        stale_marker=None,
+        fatal_error_code="",
+        progress_summary=ProgressSummary(result_subtype="success", terminal_result_text="Completed"),
+        session_log=log,
+        backup_log=tmp_path / "backup.log",
+        log_recovery=None,
+        started_epoch=0,
+        ended_epoch=1,
+    )
+
+
+def test_invalid_existing_checkpoint_blocks_terminal_success_with_completion_summary(tmp_path):
+    from prizmkit_runtime.runner_classification import classify_session
+    from prizmkit_runtime.runner_models import PromptGenerationResult
+
+    base_head = _init_git_repo_for_invalid_checkpoint_test(tmp_path)
+    artifact_dir = tmp_path / ".prizmkit" / "specs" / "001-invalid"
+    artifact_dir.mkdir(parents=True)
+    checkpoint = artifact_dir / "workflow-checkpoint.json"
+    checkpoint.write_text(json.dumps({"steps": [{"id": "S01", "skill": "done", "name": "Done", "status": "not-valid"}]}), encoding="utf-8")
+    (artifact_dir / "completion-summary.json").write_text("{}", encoding="utf-8")
+    prompt = PromptGenerationResult(output_path=tmp_path / "prompt.md", checkpoint_path=checkpoint, artifact_path=artifact_dir)
+
+    classification = classify_session(_terminal_success_result(tmp_path), project_root=tmp_path, base_head=base_head, prompt=prompt)
+
+    assert classification.session_status == "crashed"
+    assert classification.reason == "invalid_workflow_checkpoint:invalid_status"
+
+
+def test_invalid_existing_checkpoint_blocks_head_change_success(tmp_path):
+    from prizmkit_runtime.runner_classification import classify_session
+    from prizmkit_runtime.runner_models import PromptGenerationResult
+    from prizmkit_runtime.sessions import AISessionCommand, AISessionResult
+    from prizmkit_runtime.status import ProgressSummary
+
+    base_head = _init_git_repo_for_invalid_checkpoint_test(tmp_path)
+    (tmp_path / "done.txt").write_text("done", encoding="utf-8")
+    subprocess.run(["git", "add", "done.txt"], cwd=tmp_path, check=True)
+    subprocess.run(["git", "commit", "-m", "done"], cwd=tmp_path, check=True, stdout=subprocess.DEVNULL)
+    artifact_dir = tmp_path / ".prizmkit" / "specs" / "001-invalid"
+    artifact_dir.mkdir(parents=True)
+    checkpoint = artifact_dir / "workflow-checkpoint.json"
+    checkpoint.write_text(json.dumps({"steps": [{"id": "bad", "skill": "done", "name": "Done", "status": "completed"}]}), encoding="utf-8")
+    raw = AISessionResult(
+        command=AISessionCommand(("fake",), None, "claude", "stdin"),
+        exit_code=1,
+        timed_out=False,
+        termination_reason=None,
+        stale_marker=None,
+        fatal_error_code="context_overflow",
+        progress_summary=ProgressSummary(fatal_error_code="context_overflow"),
+        session_log=tmp_path / "session.log",
+        backup_log=tmp_path / "backup.log",
+        log_recovery=None,
+        started_epoch=0,
+        ended_epoch=1,
+    )
+    raw.session_log.write_text("context overflow", encoding="utf-8")
+    prompt = PromptGenerationResult(output_path=tmp_path / "prompt.md", checkpoint_path=checkpoint, artifact_path=artifact_dir)
+
+    classification = classify_session(raw, project_root=tmp_path, base_head=base_head, prompt=prompt)
+
+    assert classification.session_status == "crashed"
+    assert classification.reason == "invalid_workflow_checkpoint:malformed_step_id"
+
+
+def test_invalid_existing_checkpoint_blocks_workspace_auto_commit_success(tmp_path):
+    from prizmkit_runtime.runner_classification import classify_session
+    from prizmkit_runtime.runner_models import PromptGenerationResult
+    from prizmkit_runtime.sessions import AISessionCommand, AISessionResult
+    from prizmkit_runtime.status import ProgressSummary
+
+    base_head = _init_git_repo_for_invalid_checkpoint_test(tmp_path)
+    artifact_dir = tmp_path / ".prizmkit" / "specs" / "001-invalid"
+    artifact_dir.mkdir(parents=True)
+    checkpoint = artifact_dir / "workflow-checkpoint.json"
+    checkpoint.write_text(json.dumps({"steps": []}), encoding="utf-8")
+    (tmp_path / "dirty.txt").write_text("dirty", encoding="utf-8")
+    log = tmp_path / "session.log"
+    log.write_text("clean exit but dirty workspace", encoding="utf-8")
+    raw = AISessionResult(
+        command=AISessionCommand(("fake",), None, "claude", "stdin"),
+        exit_code=0,
+        timed_out=False,
+        termination_reason=None,
+        stale_marker=None,
+        fatal_error_code="",
+        progress_summary=ProgressSummary(result_subtype="", terminal_result_text=""),
+        session_log=log,
+        backup_log=tmp_path / "backup.log",
+        log_recovery=None,
+        started_epoch=0,
+        ended_epoch=1,
+    )
+    prompt = PromptGenerationResult(output_path=tmp_path / "prompt.md", checkpoint_path=checkpoint, artifact_path=artifact_dir)
+
+    classification = classify_session(raw, project_root=tmp_path, base_head=base_head, prompt=prompt)
+
+    assert classification.session_status == "crashed"
+    assert classification.reason == "invalid_workflow_checkpoint:empty_steps"