prizmkit 1.1.127 → 1.1.129

package/bundled/dev-pipeline/templates/sections/phase-specify-plan-full.md

@@ -71,7 +71,7 @@ Before proceeding past CP-1, verify:
 
 ### Plan Self-Check — Orchestrator Direct
 
-Do not spawn a top-level Reviewer subagent for planning analysis. Perform a bounded self-check before implementation:
+Do not delegate planning analysis. The current Main Agent performs a bounded self-check before implementation:
 
 1. Re-read only `context-snapshot.md`, `spec.md`, and `plan.md`.
 2. Cross-check spec.md and plan.md Tasks against the feature description and Verification Gates in the Task Contract.

package/bundled/dev-pipeline/templates/sections/phase0-init.md

@@ -1,6 +1,6 @@
 ### Phase 0: Project Bootstrap
 - Run `/prizmkit-init` (invoke the prizmkit-init skill)
-- Run `{{PYTHON_CMD}} {{INIT_SCRIPT_PATH}} --project-root {{PROJECT_ROOT}} --feature-id {{FEATURE_ID}} --feature-slug {{FEATURE_SLUG}}`
+- Run `{{PYTHON_CMD}} .prizmkit/dev-pipeline/scripts/init-change-artifact.py --project-root {{PROJECT_ROOT}} --feature-id {{FEATURE_ID}} --feature-slug {{FEATURE_SLUG}}`
 - **CP-0**: Verify `.prizmkit/prizm-docs/root.prizm`, `.prizmkit/config.json` exist
 
 

package/bundled/dev-pipeline/templates/sections/refactor-mission.md

@@ -4,6 +4,6 @@ You are the **refactor session orchestrator**. Execute Refactor {{REFACTOR_ID}}:
 
 **CRITICAL**: Preserve observable behavior. Do not change product behavior unless the refactor contract explicitly requires it.
 
-**Execution model**: You implement directly in the current checkout. Do not spawn a top-level Dev subagent. Do not spawn a top-level Reviewer subagent. `/prizmkit-code-review` owns its internal Reviewer Agent → filter → orchestrator fix loop.
+**Execution model**: You implement directly in the current checkout. Do not spawn a top-level Dev subagent. The current Main Agent runs `/prizmkit-code-review` directly and owns its complete review, repair, verification, and reporting loop without direct or indirect review delegation.
 
 **NON-INTERACTIVE MODE**: There is NO human on the other end. Resolve uncertainty from existing code and conservative project patterns.

package/bundled/dev-pipeline/templates/sections/refactor-phase-review.md

@@ -1,8 +1,16 @@
 ### Phase 3: Review — Code Review & Behavior Verification
 
-Run `/prizmkit-code-review` directly with `artifact_dir=.prizmkit/refactor/{{REFACTOR_ID}}/`. The skill first runs a Main-Agent self-review loop for up to ten completed rounds, then semantic risk permits independent Reviewers at limits `low=0`, `medium=1`, `high=2`. High risk starts Reviewer 2 only after Reviewer 1 produced an accepted high finding that was repaired, tested, and reconverged. Reviewer 3 is forbidden.
+Run `/prizmkit-code-review` directly with `artifact_dir=.prizmkit/refactor/{{REFACTOR_ID}}/`. The skill runs a Main-Agent review loop for up to ten completed rounds. The current Main Agent is the only Code Review executor: it reviews the complete current change, rejects unsupported candidate findings, directly repairs accepted findings, verifies repairs, and repeats until `accepted = 0` with no unresolved findings.
 
-Do not spawn a top-level Reviewer subagent. The code-review skill owns its unchanged independent Reviewer prompt, current-checkout preference, forced-worktree changed-content fallback, one bounded infrastructure retry per required Reviewer position, and append-only report progress.
+The Code Review execution boundary is absolute for this phase:
+
+- Do not delegate any review work directly or indirectly.
+- Do not invoke another review skill or review workflow.
+- Do not use a general-purpose execution unit for review or relabel delegated work as a finder, verifier, audit, compatibility review, verification, or gap sweep.
+- Direct reading, searching, editing, and testing by the current Main Agent remain allowed.
+- `review-report.md` is the only persisted review artifact; do not append separate Review Notes to another artifact.
+
+Missing required evidence, an unsafe repair, a failed verification, or failure to converge by round ten produces `NEEDS_FIXES`.
 
 **Gate Check — Final Review Result**:
 
@@ -12,9 +20,8 @@ Do not spawn a top-level Reviewer subagent. The code-review skill owns its uncha
 
 - `REVIEW_PASS` → proceed.
 - `REVIEW_NEEDS_FIXES` → log unresolved findings and proceed only when they do not block behavior preservation under explicit task policy.
-- `REVIEW_BLOCKED` → log infrastructure evidence and stop so recovery can retry.
-- `REVIEW_INCOMPLETE` or `REVIEW_INVALID` → inspect progress once, write `failure-log.md`, and stop. Do not accept stale verdict text or replace required independent review with Main-Agent self-review.
+- `REVIEW_INCOMPLETE` or `REVIEW_INVALID` → inspect progress once, write `failure-log.md`, and stop. Do not accept stale verdict text or synthesize a passing result outside the skill.
 
-Verify the review explicitly checks behavior preservation, public API compatibility, imports, tests, and rollback assumptions.
+Verify directly that the review checks behavior preservation, public API compatibility, imports, tests, and rollback assumptions.
 
-**Checkpoint update**: Set step `prizmkit-code-review` to `completed` only for an accepted valid Final Result. Do not complete the checkpoint for blocked, incomplete, or invalid reports.
+**Checkpoint update**: Set step `prizmkit-code-review` to `completed` only for an accepted valid Final Result. Do not complete the checkpoint for incomplete or invalid reports.

package/bundled/dev-pipeline/templates/sections/refactor-reminders.md

@@ -3,7 +3,7 @@
 - Use L1 skills: `/prizmkit-plan`, `/prizmkit-implement`, `/prizmkit-code-review`, `/prizmkit-retrospective`, `/prizmkit-committer`.
 - All refactor skills use `artifact_dir=.prizmkit/refactor/{{REFACTOR_ID}}/`.
 - Behavior preservation is the first priority.
-- Do not spawn top-level Dev or Reviewer subagents.
+- Do not spawn a top-level Dev implementation subagent; Code Review runs only in the current Main Agent without delegation.
 - Main orchestrator runs `/prizmkit-code-review` directly.
 - Commit with `refactor(<scope>):`, not `feat:` or `fix:`.
 - Checkpoint state is the recovery truth; skipped steps must be marked `skipped`.

package/bundled/dev-pipeline/templates/sections/subagent-timeout-recovery.md

@@ -1,6 +1,8 @@
-## Subagent Timeout Recovery
+## Normal Work Delegation Timeout Recovery
 
-If any agent times out:
+This recovery path applies only to normal work delegated outside Code Review. Code Review never uses delegated execution or this retry procedure.
+
+If a normal work execution unit times out:
 1. `{{RUNTIME_HELPER_CMD}} artifact list .prizmkit/specs/{{FEATURE_SLUG}}` — check what exists
-2. If `context-snapshot.md` exists: open recovery prompt with `"Read .prizmkit/specs/{{FEATURE_SLUG}}/context-snapshot.md for project context and any Implementation Log/Review Notes from previous agents. Run git diff HEAD to see actual code changes already made. Do NOT re-read individual source files unless the File Manifest directs you to."` + only remaining steps + `model: "lite"`
-3. Max 2 retries per phase. After 2 failures, orchestrator completes the work directly and appends a Recovery Note to context-snapshot.md.
+2. If `context-snapshot.md` exists: open a recovery prompt with `"Read .prizmkit/specs/{{FEATURE_SLUG}}/context-snapshot.md for project context and the Implementation Log. Run git diff HEAD to see actual code changes already made. Do NOT re-read individual source files unless the File Manifest directs you to."` + only remaining steps + `model: "lite"`
+3. Max 2 retries per phase. After 2 failures, the Main Agent completes the work directly and appends a Recovery Note to context-snapshot.md.

package/bundled/dev-pipeline/templates/sections/test-failure-recovery-agent.md

@@ -6,7 +6,7 @@ Use this protocol whenever the post-review PrizmKit Test gate fails. Its purpose
 
 | Class | Meaning | Required Action | May Continue? |
 |-------|---------|-----------------|---------------|
-| Baseline failure | Existed before this session | Document in Implementation Log or review notes | Yes |
+| Baseline failure | Existed before this session | Document in the Implementation Log or `failure-log.md` | Yes |
 | New regression | Introduced by this session | Fix before reporting success | No |
 | Brittle test | Test expectation/setup is wrong for the intended behavior | Fix the test or environment setup, then rerun | Only after fixed or documented as blocked |
 | Environment/tooling failure | External tool, network, install, or local environment prevents verification | Document in `failure-log.md` with impact on gates | Only if no Verification Gate is blocked |

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

@@ -1,6 +1,7 @@
 """Tests for generate-bootstrap-prompt.py core functions."""
 
 import json
+import importlib.util
 import os
 import re
 import time
@@ -25,6 +26,7 @@ from generate_bootstrap_prompt import (
     generate_checkpoint_definition,
     merge_checkpoint_state,
     load_active_agent_prompts,
+    validate_rendered,
     main as generate_bootstrap_main,
 )
 from continuation import append_continuation_handoff, checkpoint_cursor
@@ -44,6 +46,69 @@ FORBIDDEN_TEST_CMD_PLACEHOLDER = "(" + "$TEST_CMD" + ")"
 FORBIDDEN_THREE_STRIKE = "3-" + "strike"
 FORBIDDEN_TWENTY_STEP = "20-" + "step window"
 
+REVIEW_PROMPT_REQUIRED = (
+    "The current Main Agent is the only Code Review executor",
+    "Do not invoke another review skill or review workflow",
+    "finder, verifier, audit, compatibility review, verification, or gap sweep",
+    "`review-report.md` is the only persisted review artifact",
+)
+
+REVIEW_PROMPT_FORBIDDEN = (
+    "Orchestrator + Reviewer",
+    "Reviewer reviews",
+    "Reviewer subagents",
+    "Reviewer appends Review Notes",
+    "Implementation Log/Review Notes",
+    "internal Reviewer Agent",
+    "spawning Reviewer",
+    "delayed teammate/reviewer findings",
+    "Gate checks enforce Implementation Log and Review Notes",
+)
+
+
+def assert_unambiguous_main_agent_review(prompt):
+    for required in REVIEW_PROMPT_REQUIRED:
+        assert required in prompt
+    for forbidden in REVIEW_PROMPT_FORBIDDEN:
+        assert forbidden not in prompt
+
+
+PROMPT_RENDERING_FORBIDDEN = (
+    "If implementation is delegated to a Dev subagent",
+    "Document in Implementation Log or review notes",
+    "{{INIT_SCRIPT_PATH}}",
+)
+
+
+def assert_complete_prompt_rendering(prompt):
+    assert_unambiguous_main_agent_review(prompt)
+    for forbidden in PROMPT_RENDERING_FORBIDDEN:
+        assert forbidden not in prompt
+    assert re.findall(r"\{\{[A-Z][A-Z_0-9]+\}\}", prompt) == []
+
+
+def _load_recovery_prompt_module():
+    path = Path("dev-pipeline/scripts/generate-recovery-prompt.py")
+    spec = importlib.util.spec_from_file_location("generate_recovery_prompt", path)
+    module = importlib.util.module_from_spec(spec)
+    spec.loader.exec_module(module)
+    return module
+
+
+def test_rendered_prompt_validation_rejects_unresolved_control_placeholders():
+    content = (
+        "## Your Mission\n"
+        "## Execution\n"
+        "Run {{INIT_SCRIPT_PATH}}\n"
+        "## Failure Capture\n"
+    )
+
+    is_valid, warnings, errors = validate_rendered(content)
+
+    assert is_valid is False
+    assert warnings == []
+    assert errors == ["Unreplaced placeholders: {{INIT_SCRIPT_PATH}}"]
+
 
 def scoped_report_text(verdict="PASS", boundary_missing=0,
                        happy_path_only=0, completion_gate="yes",
@@ -882,9 +947,39 @@ class TestDirectOrchestratorImplementationPrompts:
             assert FORBIDDEN_TEST_CMD_PLACEHOLDER not in rendered
             assert FORBIDDEN_THREE_STRIKE not in rendered
             assert FORBIDDEN_TWENTY_STEP not in rendered
-            assert "Main-Agent self-review loop" in rendered
-            assert "low=0" in rendered and "medium=1" in rendered and "high=2" in rendered
-            assert "Reviewer 3 is forbidden" in rendered
+            assert "Main-Agent review loop" in rendered
+            assert "up to ten completed rounds" in rendered
+            assert "The current Main Agent is the only Code Review executor" in rendered
+            assert "Do not invoke another review skill or review workflow" in rendered
+            assert "low=0" not in rendered
+            assert "Reviewer 3" not in rendered
+
+    def test_all_rendered_pipeline_prompts_have_one_main_agent_review_model(self, tmp_path):
+        recovery = _load_recovery_prompt_module()
+        prompts = [
+            _render_feature_prompt(tmp_path / "feature-lite", mode="lite"),
+            _render_feature_prompt(tmp_path / "feature-standard", mode="standard"),
+            _render_feature_prompt(tmp_path / "feature-full", mode="full"),
+            _render_feature_prompt(
+                tmp_path / "feature-tier3",
+                mode="full",
+                template="bootstrap-tier3.md",
+            ),
+            _render_bugfix_prompt(tmp_path / "bugfix"),
+            _render_refactor_prompt(tmp_path / "refactor"),
+            recovery.build_bugfix_prompt(
+                {
+                    "workflow_type": "bug-fix-workflow",
+                    "phase": 5,
+                    "phase_name": "Review",
+                    "workflow_data": {"bug_id": "B-001"},
+                },
+                str(tmp_path),
+            ),
+        ]
+
+        for prompt in prompts:
+            assert_complete_prompt_rendering(prompt)
 
     def test_retained_tier3_fallback_template_uses_direct_orchestrator_implementation(self):
         rendered = Path("dev-pipeline/templates/bootstrap-tier3.md").read_text(encoding="utf-8")
@@ -1443,8 +1538,10 @@ class TestHeadlessPromptCleanupF033:
         refactor_prompt = _render_refactor_prompt(tmp_path / "refactor-mixed", platform="claude")
 
         assert "{{" + "REVIEWER_" + "SUBAGENT_PATH" + "}}" not in replacements
-        assert "Main-Agent self-review" in refactor_prompt
-        assert "Reviewer 3 is forbidden" in refactor_prompt
+        assert "Main-Agent review loop" in refactor_prompt
+        assert "The current Main Agent is the only Code Review executor" in refactor_prompt
+        assert "Do not invoke another review skill or review workflow" in refactor_prompt
+        assert "Reviewer 3" not in refactor_prompt
 
     def test_bugfix_refactor_explicit_codex_override(self, tmp_path, monkeypatch):
         from generate_bugfix_prompt import build_replacements as bugfix_build_replacements
@@ -1465,8 +1562,10 @@ class TestHeadlessPromptCleanupF033:
         refactor_prompt = _render_refactor_prompt(tmp_path / "refactor-codex", platform="claude", explicit_platform="codex")
 
         assert "{{" + "REVIEWER_" + "SUBAGENT_PATH" + "}}" not in replacements
-        assert "Main-Agent self-review" in refactor_prompt
-        assert "Reviewer 3 is forbidden" in refactor_prompt
+        assert "Main-Agent review loop" in refactor_prompt
+        assert "The current Main Agent is the only Code Review executor" in refactor_prompt
+        assert "Do not invoke another review skill or review workflow" in refactor_prompt
+        assert "Reviewer 3" not in refactor_prompt
 
 
 class TestFeatureBootstrapShellExtraction:

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

@@ -1408,7 +1408,8 @@ def test_default_no_worktree_preserves_main_worktree_branch_launch(monkeypatch,
             lifecycle_order.append("branch_create")
             branch_events.append(("create", plan.commands[1].args[2]))
         if plan.name == "branch_merge":
-            branch_events.append(("merge", plan.commands[5].args[2]))
+            merge_command = next(command for command in plan.commands if command.args[:2] == ("merge", "--ff-only"))
+            branch_events.append(("merge", merge_command.args[2]))
         return SimpleNamespace(ok=True, plan=plan, results=())
 
     def fake_start_item(*args, **kwargs):
@@ -1603,6 +1604,79 @@ def test_worktree_recovers_completed_pending_branch_from_base_session_history(mo
     assert merge_calls and cleanup_calls[-1][1] is True
 
 
+def test_success_final_commit_failure_preserves_checkout_without_merge(monkeypatch, tmp_path):
+    from prizmkit_runtime import runners
+    from prizmkit_runtime.runner_models import family_for, parse_invocation
+    from prizmkit_runtime.task_checkout import load_task_checkout
+
+    monkeypatch.setenv("DEV_BRANCH", "dev/F-001-finalize-fail")
+    paths = _make_paths(tmp_path)
+    family = family_for("feature", paths)
+    invocation = parse_invocation(family, "run", ())
+    invocation = invocation.__class__(**{**invocation.__dict__, "list_path": paths.feature_plan})
+    _install_runner_session_fakes(monkeypatch, runners)
+    merge_calls = []
+
+    monkeypatch.setattr(
+        runners,
+        "commit_final_bookkeeping",
+        lambda *args, **kwargs: SimpleNamespace(ok=False, attempted=True, committed=False),
+    )
+    monkeypatch.setattr(
+        runners,
+        "commit_task_branch_changes",
+        lambda *args, **kwargs: SimpleNamespace(ok=True, attempted=False, committed=False),
+    )
+    monkeypatch.setattr(
+        runners,
+        "run_git_plan",
+        lambda root, plan: merge_calls.append(plan.name) or SimpleNamespace(ok=True, plan=plan, results=()),
+    )
+
+    status = runners._process_item(family, invocation, "F-001", paths, initial_metadata={})
+
+    assert status == "completed"
+    assert "branch_merge" not in merge_calls
+    checkout = load_task_checkout(family.state_dir, family.task_type, "F-001")
+    assert checkout is not None and checkout.is_active
+    session_status = paths.feature_state_dir / "F-001" / "sessions" / "F-001-session-1" / "session-status.json"
+    assert json.loads(session_status.read_text(encoding="utf-8"))["status"] == "merge_conflict"
+
+
+def test_wip_commit_failure_skips_base_return(monkeypatch, tmp_path):
+    from prizmkit_runtime import runners
+    from prizmkit_runtime.runner_models import family_for, parse_invocation
+
+    monkeypatch.setenv("DEV_BRANCH", "dev/F-001-wip-fail")
+    paths = _make_paths(tmp_path)
+    family = family_for("feature", paths)
+    invocation = parse_invocation(family, "run", ())
+    invocation = invocation.__class__(**{**invocation.__dict__, "list_path": paths.feature_plan})
+    _install_runner_session_fakes(monkeypatch, runners, statuses=("crashed",), update_status="failed")
+    returns = []
+
+    monkeypatch.setattr(
+        runners,
+        "commit_wip_changes",
+        lambda *args, **kwargs: SimpleNamespace(ok=False, attempted=True, committed=False),
+    )
+    monkeypatch.setattr(
+        runners,
+        "branch_ensure_return",
+        lambda *args, **kwargs: returns.append(args) or SimpleNamespace(ok=True),
+    )
+    monkeypatch.setattr(
+        runners,
+        "run_git_plan",
+        lambda root, plan: SimpleNamespace(ok=True, plan=plan, results=()),
+    )
+
+    status = runners._process_item(family, invocation, "F-001", paths, initial_metadata={})
+
+    assert status == "failed"
+    assert returns == []
+
+
 def test_default_branch_names_keep_timestamp_for_each_runner_family(monkeypatch, tmp_path):
     from prizmkit_runtime import runners
     from prizmkit_runtime.runner_models import family_for
@@ -1752,7 +1826,7 @@ def test_foreground_runner_emits_structured_feature_progress(monkeypatch, tmp_pa
     assert "Feature: F-001 — Readable output" in captured.err
     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: full (Tier 3 — Orchestrator + Reviewer)" in captured.err
+    assert "Pipeline mode: full (Tier 3 — Main-Agent Full Guardrails)" in captured.err
     assert "Agents: 3" in captured.err
     assert "Spawning AI CLI session: F-001-session-1" in captured.err
     assert "Session result: success" in captured.err
@@ -2098,24 +2172,27 @@ def test_preflight_ready_commit_includes_tracked_state_but_excludes_untracked_st
     assert untracked_state.relative_to(tmp_path).as_posix() not in _head_files(tmp_path)
 
 
-def test_preflight_ready_commit_includes_tracked_framework_support_but_excludes_untracked_support_and_state(tmp_path):
+def test_preflight_ready_commit_excludes_tracked_framework_manifest_but_preserves_other_tracked_support(tmp_path):
     from prizmkit_runtime.runner_bookkeeping import commit_preflight_ready_changes
 
     _init_bookkeeping_repo(tmp_path)
+    manifest_path = tmp_path / ".prizmkit" / "manifest.json"
     tracked_support_paths = [
-        tmp_path / ".prizmkit" / "manifest.json",
         tmp_path / ".prizmkit" / "dev-pipeline" / "prizmkit_runtime" / "runners.py",
         tmp_path / ".claude" / "commands" / "prizmkit.md",
     ]
+    manifest_path.parent.mkdir(parents=True, exist_ok=True)
+    manifest_path.write_text('{"version":"1.1.127"}\n', encoding="utf-8")
     for path in tracked_support_paths:
         path.parent.mkdir(parents=True, exist_ok=True)
         path.write_text("installed\n", encoding="utf-8")
     subprocess.run(
-        ["git", "add", *(path.relative_to(tmp_path).as_posix() for path in tracked_support_paths)],
+        ["git", "add", manifest_path.relative_to(tmp_path).as_posix(), *(path.relative_to(tmp_path).as_posix() for path in tracked_support_paths)],
         cwd=tmp_path,
         check=True,
     )
     subprocess.run(["git", "commit", "-m", "track support"], cwd=tmp_path, check=True, stdout=subprocess.DEVNULL)
+    manifest_path.write_text('{"version":"1.1.128"}\n', encoding="utf-8")
     for path in tracked_support_paths:
         path.write_text("upgraded\n", encoding="utf-8")
 
@@ -2129,19 +2206,23 @@ def test_preflight_ready_commit_includes_tracked_framework_support_but_excludes_
         path.write_text("transient\n", encoding="utf-8")
 
     result = commit_preflight_ready_changes(tmp_path, "B-001")
-    committed_files = _head_files(tmp_path)
 
-    assert result.committed is True
-    assert {path.relative_to(tmp_path).as_posix() for path in tracked_support_paths}.issubset(committed_files)
-    assert not {path.relative_to(tmp_path).as_posix() for path in untracked_support_paths}.intersection(committed_files)
-    assert _head_message(tmp_path) == "chore: ready for B-001"
+    assert result.attempted is True
+    assert result.committed is False
+    assert result.result is not None
+    assert "installer-owned .prizmkit/manifest.json" in result.result.stderr
+    assert _head_message(tmp_path) == "track support"
     assert subprocess.run(
         ["git", "status", "--short", "--untracked-files=no"],
         cwd=tmp_path,
         check=True,
         stdout=subprocess.PIPE,
         text=True,
-    ).stdout == ""
+    ).stdout.splitlines() == [
+        " M .claude/commands/prizmkit.md",
+        " M .prizmkit/dev-pipeline/prizmkit_runtime/runners.py",
+        " M .prizmkit/manifest.json",
+    ]
 
 
 def test_preflight_ready_commit_excludes_untracked_support_assets_and_state(tmp_path):
@@ -2176,6 +2257,53 @@ def test_preflight_ready_commit_excludes_untracked_support_assets_and_state(tmp_
     assert ".claude" in hidden_status or ".prizmkit/state" in hidden_status
 
 
+def test_preflight_ready_commit_rejects_unmerged_index(tmp_path):
+    from prizmkit_runtime.runner_bookkeeping import commit_preflight_ready_changes
+
+    _init_bookkeeping_repo(tmp_path)
+    subprocess.run(["git", "checkout", "-b", "other"], cwd=tmp_path, check=True, stdout=subprocess.DEVNULL)
+    (tmp_path / "staged.txt").write_text("other\n", encoding="utf-8")
+    subprocess.run(["git", "commit", "-am", "other"], cwd=tmp_path, check=True, stdout=subprocess.DEVNULL)
+    subprocess.run(["git", "checkout", "main"], cwd=tmp_path, check=True, stdout=subprocess.DEVNULL)
+    (tmp_path / "staged.txt").write_text("main\n", encoding="utf-8")
+    subprocess.run(["git", "commit", "-am", "main"], cwd=tmp_path, check=True, stdout=subprocess.DEVNULL)
+    subprocess.run(["git", "merge", "other"], cwd=tmp_path, check=False, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL)
+    before_count = _head_commit_count(tmp_path)
+
+    result = commit_preflight_ready_changes(tmp_path, "F-126")
+
+    assert result.attempted is True
+    assert result.committed is False
+    assert result.result is not None
+    assert "unmerged Git paths" in result.result.stderr
+    assert _head_commit_count(tmp_path) == before_count
+
+
+def test_preflight_ready_commit_rejects_conflict_markers_and_malformed_manifest(tmp_path):
+    from prizmkit_runtime.runner_bookkeeping import commit_preflight_ready_changes
+
+    _init_bookkeeping_repo(tmp_path)
+    (tmp_path / "unstaged.txt").write_text("<<<<<<< Updated upstream\ncurrent\n=======\nstashed\n>>>>>>> Stashed changes\n", encoding="utf-8")
+
+    marker_result = commit_preflight_ready_changes(tmp_path, "F-127")
+
+    assert marker_result.committed is False
+    assert marker_result.result is not None
+    assert "unresolved conflict markers" in marker_result.result.stderr
+
+    (tmp_path / "unstaged.txt").write_text("safe\n", encoding="utf-8")
+    manifest = tmp_path / ".prizmkit" / "manifest.json"
+    manifest.parent.mkdir(parents=True, exist_ok=True)
+    manifest.write_text("<<<<<<< Updated upstream\n{}\n", encoding="utf-8")
+
+    manifest_result = commit_preflight_ready_changes(tmp_path, "F-128")
+
+    assert manifest_result.committed is False
+    assert manifest_result.result is not None
+    assert ".prizmkit/manifest.json" in manifest_result.result.stderr
+    assert "valid JSON" in manifest_result.result.stderr
+
+
 def test_preflight_ready_commit_noops_when_workspace_has_no_visible_dirty_content(tmp_path):
     from prizmkit_runtime.runner_bookkeeping import commit_preflight_ready_changes
 

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

@@ -100,11 +100,17 @@ class TestArtifactAndTextHelpers:
         report = tmp_path / "review-report.md"
         report.write_text(
             "## Final Result\n- Verdict: PASS\n\n"
-            "## Final Result\n- Verdict: BLOCKED\n",
+            "## Final Result\n- Verdict: NEEDS_FIXES\n",
             encoding="utf-8",
         )
-        blocked = run_helper("text", "final-verdict", str(report))
-        assert first_line(blocked) == "REVIEW_BLOCKED"
+        needs_fixes = run_helper("text", "final-verdict", str(report))
+        assert needs_fixes.returncode == 0
+        assert first_line(needs_fixes) == "REVIEW_NEEDS_FIXES"
+
+        report.write_text("## Final Result\n- Verdict: BLOCKED\n", encoding="utf-8")
+        retired = run_helper("text", "final-verdict", str(report))
+        assert retired.returncode != 0
+        assert first_line(retired) == "REVIEW_INVALID"
 
         report.write_text("## Final Result\n- Verdict: UNKNOWN\n", encoding="utf-8")
         invalid = run_helper("text", "final-verdict", str(report))