11 deleted/transferred tests because wf_api has some

2026-06-02 03:53:05 +07:00
parent 69de023a89
commit 4e28c8db5d
8 changed files with 769 additions and 333 deletions
@@ -442,6 +442,26 @@ Do not move `wf_mcp.shared.pagination` in this slice. It is still used by proxy
 tool search/listing code, so treating it as dead workflow-surface debt would be
 incorrect.

+### Post-5 Helper Cleanup: Workflow Surface Test Thinning
+
+Concrete plan:
+
+```text
+docs/superpowers/plans/2026-06-02-wf-mcp-workflow-surface-test-thinning.md
+```
+
+Intent:
+
+```text
+wf_api tests                 = canonical application behavior tests
+wf_mcp.workflow_surface tests = adapter/schema/live-source/integration smoke tests
+```
+
+Do not replace stronger workflow-surface integration tests with weaker unit
+tests. Only remove a handler test when an equal-or-stronger `wf_api` test exists
+and at least one handler-level smoke/delegation test still protects the adapter
+path.
+
 ### Candidate Moves

 ```text
@@ -0,0 +1,69 @@
+# wf_mcp Workflow Surface Test Thinning Ledger
+
+This ledger records every `wf_mcp.workflow_surface` test removed or kept during
+the thinning pass. Do not delete a test unless the `replacement` column points
+to equal-or-stronger coverage.
+
+| Test | Decision | Replacement / Reason |
+| --- | --- | --- |
+| test_artifacts::test_workflow_surface_lists_artifact_catalog_entries | keep | Handler adapter smoke test for compact artifact listing; reduced to minimal assertions. |
+| test_artifacts::test_workflow_surface_pages_and_filters_artifact_catalog_entries | remove | Covered by WorkflowArtifactApi list tests plus wf_api.listing pagination tests; handler keeps one list_artifacts smoke test. |
+| test_capabilities::test_workflow_surface_lists_planner_visible_capabilities | keep | Handler list smoke test; reduced to minimal adapter-path assertions. |
+| test_capabilities::test_workflow_surface_filters_stdlib_capabilities_by_source | remove | Covered by WorkflowCapabilityApi source/query filtering; handler list smoke remains. |
+| test_capabilities::test_workflow_surface_call_capability_returns_structured_error | keep | Protected: MCP/service behavior — structured error on capability call failure. |
+| test_capabilities::test_workflow_surface_lists_saved_wrapper_capabilities | remove | Covered by WorkflowCapabilityApi saved wrapper list tests. |
+| test_capabilities::test_workflow_surface_inspects_one_capability | keep | Handler inspect smoke test; ensures adapter path works for single capability inspection. |
+| test_capabilities::test_workflow_surface_inspect_capability_includes_wrapper_hints | keep | Protected: MCP/service behavior — wrapper hints detail through handler. |
+| test_capabilities::test_workflow_surface_inspects_saved_wrapper_capability | remove | Covered by WorkflowCapabilityApi saved wrapper inspect tests. |
+| test_capabilities::test_workflow_surface_does_not_auto_map_raw_mcp_content_blocks | keep | Protected: MCP content-block mapping behavior. |
+| test_deployments::test_workflow_surface_validates_deployment_dependencies | keep | Handler-level dependency validation has stronger next-action assertions than wf_api. |
+| test_deployments::test_workflow_surface_validate_deployment_live_check_is_opt_in | keep | Protected: MCP service adapter — live check is opt-in. |
+| test_deployments::test_workflow_surface_validate_deployment_live_check_reports_unreachable_source | keep | Protected: MCP service adapter — unreachable source reporting. |
+| test_deployments::test_workflow_surface_validate_deployment_live_check_reports_missing_connection | keep | Protected: MCP service adapter — missing connection reporting. |
+| test_deployments::test_workflow_surface_records_artifact_and_deployment_save_events | keep | Protected: service event recording. |
+| test_deployments::test_workflow_surface_save_deployment_accepts_deployment_id_alias | keep | Protected: request alias normalization. |
+| test_deployments::test_workflow_surface_deletes_deployment | keep | Protected: delete event recording. |
+| test_deployments::test_workflow_surface_save_deployment_rejects_id_and_deployment_id | keep | Protected: XOR validation on id/deployment_id. |
+| test_deployments::test_workflow_surface_lists_compact_deployment_summaries_and_inspects_detail | keep | Protected: compact-vs-detail response shape. |
+| test_drafts::test_workflow_surface_validates_draft_without_saving | remove | Covered by WorkflowDraftApi.validate_draft via test_delegation_smoke_validate_draft_equivalence. |
+| test_drafts::test_workflow_surface_rejects_unknown_draft_route_outcome_when_spec_is_known | keep | Live outcome lookup through handler; not duplicated in wf_api. |
+| test_drafts::test_workflow_surface_creates_artifact_from_draft_with_binding_suggestions | keep | Binding suggestions and artifact persistence through handler. |
+| test_drafts::test_workflow_surface_draft_artifact_requires_std_self_binding | keep | binding_missing diagnostic through handler/deployment integration. |
+| test_drafts::test_workflow_surface_patches_draft_without_saving | remove | Covered by WorkflowDraftApi.patch_draft. |
+| test_drafts::test_workflow_surface_creates_and_gets_draft_workspace | remove | Covered by WorkflowDraftApi.create_draft_workspace and get_draft_workspace assertions in tests/wf_api/test_drafts_service.py. |
+| test_drafts::test_workflow_surface_lists_draft_workspaces | remove | Covered by tests/wf_api/test_drafts_service.py::test_list_draft_workspaces_returns_sorted_summaries_without_drafts. |
+| test_drafts::test_workflow_surface_deletes_draft_workspace | remove | Covered by tests/wf_api/test_drafts_service.py::test_delete_draft_workspace_is_idempotent. |
+| test_drafts::test_workflow_surface_patch_helpers_update_draft_workspace | remove | Covered by tests/wf_api/test_drafts_service.py::test_draft_workspace_patch_helpers_update_revision_and_bindings. |
+| test_drafts::test_workflow_surface_validates_draft_workspace_with_live_outcomes | keep | Live outcome lookup through handler/service stack. |
+| test_drafts::test_workflow_surface_patches_draft_workspace_by_revision | remove | Covered by WorkflowDraftApi.patch_draft_workspace. |
+| test_drafts::test_workflow_surface_creates_minimal_draft_workspace_with_error_route | keep | MCP request model parsing and error route generation. |
+| test_drafts::test_workflow_surface_minimal_draft_honors_explicit_error_message_source | keep | Explicit error_message_source handling. |
+| test_drafts::test_minimal_draft_request_accepts_structural_error_message_source | keep | MCP Pydantic model validation (CreateMinimalDraftWorkspaceRequest). |
+| test_drafts::test_workflow_surface_accepts_canonical_bindings_for_minimal_workspace | keep | Canonical InputPathBinding/OutputBinding through handler. |
+| test_drafts::test_workflow_surface_creates_draft_workspace_from_capability_hints | keep | Wrapper hints and next-actions through handler. |
+| test_drafts::test_workflow_surface_creates_artifact_from_workspace | keep | Artifact persistence with schema snapshots through handler. |
+| test_drafts::test_workflow_surface_workspace_artifact_infers_raw_concrete_dependency | keep | Source dependency inference through handler. |
+| test_drafts::test_workflow_surface_creates_wrapper_from_workspace | keep | Wrapper creation through handler. |
+| test_drafts::test_workflow_surface_low_confidence_draft_returns_patch_guidance | keep | Next-action guidance with patch examples through handler. |
+| test_runs::test_raw_workflow_plan_uses_core_step_and_edge_models | keep | DEVIATION: Plan said wf_api covers this, but test_raw_workflow_plan_extraction.py only tests imports (canonical, compat, identity). The surface test is the only one exercising actual RawWorkflowPlan step/edge model parsing. Kept to avoid coverage gap. |
+| test_runs::test_workflow_surface_runs_non_interrupting_deployment | keep | Persisted run records, trace slicing, response model, next-actions. |
+| test_runs::test_workflow_surface_failed_deployment_exposes_error_on_run_and_inspect | keep | Failed run error exposure and inspect_run. |
+| test_runs::test_workflow_surface_run_deployment_can_include_trace_detail | keep | Protected: MCP TraceRange, RunDeploymentResult model validation. |
+| test_runs::test_workflow_surface_run_deployment_can_read_empty_trace_range | keep | Protected: empty trace range behavior. |
+| test_runs::test_workflow_surface_runs_deployment_with_bound_node_spec_dependency | keep | Protected: logical source binding. |
+| test_runs::test_workflow_surface_runs_artifact_created_from_concrete_node_ref | keep | Protected: concrete node ref artifact creation and run. |
+| test_runs::test_workflow_surface_detects_drift_from_saved_node_spec_snapshot | keep | Protected: schema drift detection. |
+| test_runs::test_workflow_surface_runs_deployment_with_bound_reducer_dependency | keep | Protected: reducer dependency integration. |
+| test_wrappers::test_workflow_surface_creates_wrapper_artifact_from_plan | keep | Handler integration: wrapper artifact creation from plan. |
+| test_wrappers::test_workflow_surface_creates_artifact_with_logical_node_refs | keep | Handler integration: logical node ref resolution. |
+| test_wrappers::test_workflow_surface_calls_saved_wrapper_artifact | keep | Handler integration: saved wrapper direct call. |
+| test_wrappers::test_workflow_surface_calls_live_node_spec_with_self_describing_response | keep | Handler integration: live node spec call. |
+| test_wrappers::test_workflow_surface_calls_saved_wrapper_artifact_with_deployment_bindings | keep | Protected: saved wrapper call with deployment bindings. |
+| test_next_actions::test_next_actions_from_high_confidence_wrapper_hints_can_validate | keep | NextActions model unit test. |
+| test_next_actions::test_next_actions_from_low_confidence_wrapper_hints_can_patch | keep | NextActions model unit test. |
+| test_next_actions::test_next_actions_from_runnable_deployment_recommends_run | keep | NextActions model unit test. |
+| test_next_actions::test_next_actions_from_unrunnable_deployment_recommends_validation_retry | keep | NextActions model unit test. |
+| test_next_actions::test_next_actions_from_completed_run_has_no_required_next_tool | keep | NextActions model unit test. |
+| test_next_actions::test_next_actions_from_failed_run_recommends_bounded_trace | keep | NextActions model unit test. |
+| test_next_actions::test_next_actions_from_interrupted_run_recommends_resume | keep | NextActions model unit test. |
+| test_next_actions::test_workflow_surface_next_actions_shim_reexports_canonical_model | keep | Shim re-export verification. |
@@ -0,0 +1,538 @@
+# wf_mcp Workflow Surface Test Thinning Implementation Plan
+
+> **For agentic workers:** REQUIRED SUB-SKILL: Use superpowers:subagent-driven-development (recommended) or superpowers:executing-plans to implement this plan task-by-task. Steps use checkbox (`- [ ]`) syntax for tracking.
+
+**Goal:** Reduce duplicated `wf_mcp.workflow_surface` behavior tests now covered by `wf_api` tests while preserving adapter, schema, live-source, run, wrapper, and end-to-end coverage.
+
+**Architecture:** `wf_api` tests are the canonical behavior tests for application-service logic. `wf_mcp.workflow_surface` tests should prove that `WorkflowSurfaceHandlers` exposes that behavior through the MCP-owned adapter boundary and should keep any test that exercises MCP-specific request models, live checks, service event recording, tool schema, or realistic integration paths.
+
+**Tech Stack:** Python 3.14, pytest, `wf_api`, `wf_mcp.workflow_surface`, `WorkflowSurfaceHandlers`, ruff, basedpyright.
+
+---
+
+## Coverage Policy
+
+Use this rule for every deletion:
+
+```text
+Only remove a wf_mcp.workflow_surface test when an equal-or-stronger wf_api test
+already covers the behavior and a smaller handler smoke/delegation test still
+proves the adapter path.
+```
+
+Keep tests that cover:
+
+- MCP request/response Pydantic models such as `TraceRange`, `RunDeploymentResult`, and `CreateMinimalDraftWorkspaceRequest`.
+- live source checks through `WfMcpService` adapters.
+- service event recording.
+- handler-to-service wiring.
+- saved wrapper calls with deployment bindings.
+- source binding, schema drift, reducer dependency, subgraph, or durable run integration.
+- next-action guidance generated through real handler operations.
+
+Thin tests that only repeat:
+
+- list filtering/pagination details already covered by `WorkflowCapabilityApi` or `WorkflowArtifactApi`.
+- basic inspect/list payload details already covered by `wf_api` domain tests.
+- basic draft workspace CRUD details already covered by `WorkflowDraftApi`.
+- basic artifact/deployment CRUD details already covered by `WorkflowArtifactApi` or `WorkflowDeploymentApi`.
+
+## File Map
+
+| File | Planned role |
+| --- | --- |
+| `tests/wf_mcp/workflow_surface/test_artifacts.py` | Thin to one handler adapter smoke test for compact artifact listing. |
+| `tests/wf_mcp/workflow_surface/test_capabilities.py` | Thin list/filter/inspect duplicates; keep wrapper-hints, MCP content-block, direct call error, and saved wrapper coverage if not stronger in `wf_api`. |
+| `tests/wf_mcp/workflow_surface/test_deployments.py` | Keep live-check, events, alias/XOR, delete event, and compact-vs-detail tests. Maybe remove only duplicate dependency validation if covered by `wf_api`. |
+| `tests/wf_mcp/workflow_surface/test_drafts.py` | Keep most tests; remove only the simplest validate/patch/list CRUD duplicates if `wf_api` has equal coverage. |
+| `tests/wf_mcp/workflow_surface/test_runs.py` | Keep run/trace/model tests. Do not thin in first pass except the raw plan model extraction test if duplicated in `tests/wf_api/test_raw_workflow_plan_extraction.py`. |
+| `tests/wf_mcp/workflow_surface/test_wrappers.py` | Keep all tests in first pass; these are handler integration paths and direct capability REPL behavior. |
+| `tests/wf_api/*` | Do not weaken. Add missing behavior tests here before removing handler duplicates. |
+
+---
+
+## Task 1: Build A Deletion Ledger Before Editing
+
+**Files:**
+- Create: `docs/superpowers/plans/2026-06-02-wf-mcp-workflow-surface-test-thinning-ledger.md`
+
+- [ ] **Step 1: Create the ledger file**
+
+Create `docs/superpowers/plans/2026-06-02-wf-mcp-workflow-surface-test-thinning-ledger.md`:
+
+```markdown
+# wf_mcp Workflow Surface Test Thinning Ledger
+
+This ledger records every `wf_mcp.workflow_surface` test removed or kept during
+the thinning pass. Do not delete a test unless the `replacement` column points
+to equal-or-stronger coverage.
+
+| Test | Decision | Replacement / Reason |
+| --- | --- | --- |
+```
+
+- [ ] **Step 2: Populate the initial ledger with all workflow-surface tests**
+
+Run:
+
+```bash
+rg -n "^def test_" tests/wf_mcp/workflow_surface
+```
+
+Append each test name to the ledger with `Decision` set to `unclassified`.
+
+Expected: the ledger contains every test from:
+
+```text
+tests/wf_mcp/workflow_surface/test_artifacts.py
+tests/wf_mcp/workflow_surface/test_capabilities.py
+tests/wf_mcp/workflow_surface/test_deployments.py
+tests/wf_mcp/workflow_surface/test_drafts.py
+tests/wf_mcp/workflow_surface/test_runs.py
+tests/wf_mcp/workflow_surface/test_wrappers.py
+tests/wf_mcp/workflow_surface/test_next_actions.py
+```
+
+- [ ] **Step 3: Mark protected tests**
+
+Mark these as `keep` unless a later task explicitly adds stronger coverage:
+
+```text
+tests/wf_mcp/workflow_surface/test_capabilities.py::test_workflow_surface_call_capability_returns_structured_error
+tests/wf_mcp/workflow_surface/test_capabilities.py::test_workflow_surface_inspect_capability_includes_wrapper_hints
+tests/wf_mcp/workflow_surface/test_capabilities.py::test_workflow_surface_does_not_auto_map_raw_mcp_content_blocks
+tests/wf_mcp/workflow_surface/test_deployments.py::test_workflow_surface_validate_deployment_live_check_is_opt_in
+tests/wf_mcp/workflow_surface/test_deployments.py::test_workflow_surface_validate_deployment_live_check_reports_unreachable_source
+tests/wf_mcp/workflow_surface/test_deployments.py::test_workflow_surface_validate_deployment_live_check_reports_missing_connection
+tests/wf_mcp/workflow_surface/test_deployments.py::test_workflow_surface_records_artifact_and_deployment_save_events
+tests/wf_mcp/workflow_surface/test_deployments.py::test_workflow_surface_save_deployment_accepts_deployment_id_alias
+tests/wf_mcp/workflow_surface/test_deployments.py::test_workflow_surface_save_deployment_rejects_id_and_deployment_id
+tests/wf_mcp/workflow_surface/test_runs.py::test_workflow_surface_run_deployment_can_include_trace_detail
+tests/wf_mcp/workflow_surface/test_runs.py::test_workflow_surface_run_deployment_can_read_empty_trace_range
+tests/wf_mcp/workflow_surface/test_runs.py::test_workflow_surface_runs_deployment_with_bound_node_spec_dependency
+tests/wf_mcp/workflow_surface/test_runs.py::test_workflow_surface_runs_artifact_created_from_concrete_node_ref
+tests/wf_mcp/workflow_surface/test_runs.py::test_workflow_surface_detects_drift_from_saved_node_spec_snapshot
+tests/wf_mcp/workflow_surface/test_runs.py::test_workflow_surface_runs_deployment_with_bound_reducer_dependency
+tests/wf_mcp/workflow_surface/test_wrappers.py
+```
+
+Reason: these cover adapter wiring, MCP/service behavior, direct wrapper calls,
+trace schema behavior, binding logic, reducer dependencies, or important
+integration seams.
+
+---
+
+## Task 2: Thin Artifact Listing Duplicates
+
+**Files:**
+- Modify: `tests/wf_mcp/workflow_surface/test_artifacts.py`
+- Modify: `docs/superpowers/plans/2026-06-02-wf-mcp-workflow-surface-test-thinning-ledger.md`
+- Test: `tests/wf_api/test_artifact_api.py`, `tests/wf_mcp/workflow_surface/test_artifacts.py`
+
+- [ ] **Step 1: Confirm wf_api artifact coverage**
+
+Run:
+
+```bash
+rg -n "list_artifacts|pagination|kind|query|plan\" not in" tests/wf_api/test_artifact_api.py tests/wf_api/test_listing.py
+```
+
+Expected: `wf_api` coverage exists for empty list payload, compact artifact rows,
+query/kind behavior, and pagination helper shape. If it does not, add the
+missing assertion to `tests/wf_api/test_artifact_api.py` before deleting any
+handler test.
+
+- [ ] **Step 2: Keep one handler smoke test**
+
+Keep `test_workflow_surface_lists_artifact_catalog_entries` and reduce it only
+if needed to these adapter-boundary assertions:
+
+```python
+def test_workflow_surface_lists_artifact_catalog_entries() -> None:
+    artifact_store = FileWorkflowArtifactStore(local_temp_root() / "surface_artifacts")
+    artifact_store.save_artifact(artifact())
+    h = handlers(artifact_store)
+
+    payload = asyncio.run(h.list_artifacts())
+
+    nodes = payload["nodes"]
+    assert payload["total"] == 1
+    assert payload["next_cursor"] is None
+    assert nodes[0]["name"] == "workflow.summarize_docs.v1"
+    assert nodes[0]["artifact_id"] == "summarize_docs"
+    assert "plan" not in nodes[0]
+```
+
+- [ ] **Step 3: Remove duplicate pagination/filter handler test**
+
+Delete:
+
+```text
+test_workflow_surface_pages_and_filters_artifact_catalog_entries
+```
+
+Mark it in the ledger:
+
+```text
+remove | Covered by WorkflowArtifactApi list tests plus wf_api.listing pagination tests; handler keeps one list_artifacts smoke test.
+```
+
+- [ ] **Step 4: Run artifact tests**
+
+Run:
+
+```bash
+uv run pytest tests/wf_api/test_artifact_api.py tests/wf_api/test_listing.py tests/wf_mcp/workflow_surface/test_artifacts.py -q
+```
+
+Expected: pass.
+
+---
+
+## Task 3: Thin Capability List/Inspect Duplicates Conservatively
+
+**Files:**
+- Modify: `tests/wf_mcp/workflow_surface/test_capabilities.py`
+- Modify: `docs/superpowers/plans/2026-06-02-wf-mcp-workflow-surface-test-thinning-ledger.md`
+- Test: `tests/wf_api/test_capability_api.py`, `tests/wf_mcp/workflow_surface/test_capabilities.py`
+
+- [ ] **Step 1: Confirm wf_api capability coverage**
+
+Run:
+
+```bash
+rg -n "list_capabilities|inspect_capability|saved_wrapper|wrapper_hints|runtime_error|content blocks" tests/wf_api/test_capability_api.py tests/wf_mcp/workflow_surface/test_capabilities.py
+```
+
+Expected: `wf_api` covers planner-visible listing, source filtering, unknown
+inspect, saved wrapper list/inspect, and direct handler delegation smoke.
+
+- [ ] **Step 2: Keep one handler list smoke test**
+
+Keep `test_workflow_surface_lists_planner_visible_capabilities`, but keep it
+compact. It should assert adapter path and summary shape only:
+
+```python
+def test_workflow_surface_lists_planner_visible_capabilities() -> None:
+    h = handlers(FileWorkflowArtifactStore(local_temp_root() / "surface_caps"))
+
+    payload = asyncio.run(h.list_capabilities(limit=2))
+    first = payload["capabilities"][0]
+
+    assert len(payload["capabilities"]) == 2
+    assert payload["total"] >= 2
+    assert payload["next_cursor"] == "2"
+    assert first["kind"] == "node_spec"
+    assert "input_schema" not in first
+```
+
+- [ ] **Step 3: Remove duplicate source-filter test if wf_api has it**
+
+Delete:
+
+```text
+test_workflow_surface_filters_stdlib_capabilities_by_source
+```
+
+Only delete it if `tests/wf_api/test_capability_api.py` contains:
+
+```text
+test_list_capabilities_filters_by_source
+```
+
+Mark the ledger:
+
+```text
+remove | Covered by WorkflowCapabilityApi source/query filtering; handler list smoke remains.
+```
+
+- [ ] **Step 4: Remove duplicate saved-wrapper list/inspect tests if wf_api has them**
+
+Delete these only if `tests/wf_api/test_capability_api.py` has equivalent saved
+wrapper list and inspect tests:
+
+```text
+test_workflow_surface_lists_saved_wrapper_capabilities
+test_workflow_surface_inspects_saved_wrapper_capability
+```
+
+Mark the ledger:
+
+```text
+remove | Covered by WorkflowCapabilityApi saved wrapper list/inspect tests.
+```
+
+- [ ] **Step 5: Keep MCP-specific and guidance-sensitive capability tests**
+
+Keep these tests unchanged:
+
+```text
+test_workflow_surface_call_capability_returns_structured_error
+test_workflow_surface_inspect_capability_includes_wrapper_hints
+test_workflow_surface_does_not_auto_map_raw_mcp_content_blocks
+```
+
+Do not remove `test_workflow_surface_inspects_one_capability` unless there is
+still another handler-level inspect smoke test after this task.
+
+- [ ] **Step 6: Run capability tests**
+
+Run:
+
+```bash
+uv run pytest tests/wf_api/test_capability_api.py tests/wf_mcp/workflow_surface/test_capabilities.py -q
+```
+
+Expected: pass.
+
+---
+
+## Task 4: Thin Deployment Tests Only Where Purely Duplicated
+
+**Files:**
+- Modify: `tests/wf_mcp/workflow_surface/test_deployments.py`
+- Modify: `docs/superpowers/plans/2026-06-02-wf-mcp-workflow-surface-test-thinning-ledger.md`
+- Test: `tests/wf_api/test_deployment_api.py`, `tests/wf_mcp/workflow_surface/test_deployments.py`
+
+- [ ] **Step 1: Keep MCP live-check and event tests**
+
+Do not delete:
+
+```text
+test_workflow_surface_validate_deployment_live_check_is_opt_in
+test_workflow_surface_validate_deployment_live_check_reports_unreachable_source
+test_workflow_surface_validate_deployment_live_check_reports_missing_connection
+test_workflow_surface_records_artifact_and_deployment_save_events
+test_workflow_surface_save_deployment_accepts_deployment_id_alias
+test_workflow_surface_deletes_deployment
+test_workflow_surface_save_deployment_rejects_id_and_deployment_id
+test_workflow_surface_lists_compact_deployment_summaries_and_inspects_detail
+```
+
+Reason: these exercise MCP service adapters, event recording, request alias
+normalization, and compact-vs-detail response shape.
+
+- [ ] **Step 2: Evaluate basic dependency validation duplicate**
+
+Check whether `tests/wf_api/test_deployment_api.py` covers dependency validation
+next-actions for an unrunnable deployment:
+
+```bash
+rg -n "source_missing|binding_missing|next_actions|unrunnable" tests/wf_api/test_deployment_api.py
+```
+
+If the `wf_api` test has equal next-action assertions, delete:
+
+```text
+test_workflow_surface_validates_deployment_dependencies
+```
+
+If it does not, keep the handler test and mark it:
+
+```text
+keep | Handler-level dependency validation still has stronger next-action assertions than wf_api.
+```
+
+- [ ] **Step 3: Run deployment tests**
+
+Run:
+
+```bash
+uv run pytest tests/wf_api/test_deployment_api.py tests/wf_mcp/workflow_surface/test_deployments.py -q
+```
+
+Expected: pass.
+
+---
+
+## Task 5: Thin Draft Tests With High Caution
+
+**Files:**
+- Modify: `tests/wf_mcp/workflow_surface/test_drafts.py`
+- Modify: `docs/superpowers/plans/2026-06-02-wf-mcp-workflow-surface-test-thinning-ledger.md`
+- Test: `tests/wf_api/test_drafts_service.py`, `tests/wf_mcp/workflow_surface/test_drafts.py`
+
+- [ ] **Step 1: Identify pure draft API duplicates**
+
+Run:
+
+```bash
+rg -n "validate_draft|patch_draft|list_draft_workspaces|delete_draft_workspace|create_minimal_draft_workspace|create_draft_workspace_from_capability|create_artifact_from_workspace|create_wrapper_from_workspace" tests/wf_api/test_drafts_service.py tests/wf_api/test_artifact_api.py tests/wf_api/test_capability_api.py tests/wf_mcp/workflow_surface/test_drafts.py
+```
+
+Expected: most simple draft workspace CRUD and patch helper behavior exists in
+`tests/wf_api/test_drafts_service.py`.
+
+- [ ] **Step 2: Remove only pure CRUD duplicates**
+
+Candidates for removal if `wf_api` tests cover equal behavior:
+
+```text
+test_workflow_surface_validates_draft_without_saving
+test_workflow_surface_patches_draft_without_saving
+test_workflow_surface_creates_and_gets_draft_workspace
+test_workflow_surface_lists_draft_workspaces
+test_workflow_surface_deletes_draft_workspace
+test_workflow_surface_patch_helpers_update_draft_workspace
+test_workflow_surface_patches_draft_workspace_by_revision
+```
+
+For each removed test, add a ledger row with the exact `wf_api` replacement test.
+
+- [ ] **Step 3: Keep guidance, model, and artifact integration tests**
+
+Do not delete these in this pass:
+
+```text
+test_workflow_surface_rejects_unknown_draft_route_outcome_when_spec_is_known
+test_workflow_surface_creates_artifact_from_draft_with_binding_suggestions
+test_workflow_surface_draft_artifact_requires_std_self_binding
+test_workflow_surface_validates_draft_workspace_with_live_outcomes
+test_workflow_surface_creates_minimal_draft_workspace_with_error_route
+test_workflow_surface_minimal_draft_honors_explicit_error_message_source
+test_minimal_draft_request_accepts_structural_error_message_source
+test_workflow_surface_accepts_canonical_bindings_for_minimal_workspace
+test_workflow_surface_creates_draft_workspace_from_capability_hints
+test_workflow_surface_creates_artifact_from_workspace
+test_workflow_surface_workspace_artifact_infers_raw_concrete_dependency
+test_workflow_surface_creates_wrapper_from_workspace
+test_workflow_surface_low_confidence_draft_returns_patch_guidance
+```
+
+Reason: these cover live outcome lookup, request model parsing, wrapper hints,
+next actions, artifact persistence, and source dependency inference.
+
+- [ ] **Step 4: Run draft tests**
+
+Run:
+
+```bash
+uv run pytest tests/wf_api/test_drafts_service.py tests/wf_api/test_artifact_api.py tests/wf_api/test_capability_api.py tests/wf_mcp/workflow_surface/test_drafts.py -q
+```
+
+Expected: pass.
+
+---
+
+## Task 6: Keep Run And Wrapper Tests Mostly Intact
+
+**Files:**
+- Modify: `tests/wf_mcp/workflow_surface/test_runs.py`
+- Modify: `docs/superpowers/plans/2026-06-02-wf-mcp-workflow-surface-test-thinning-ledger.md`
+- Test: `tests/wf_api/test_run_api.py`, `tests/wf_mcp/workflow_surface/test_runs.py`, `tests/wf_mcp/workflow_surface/test_wrappers.py`
+
+- [ ] **Step 1: Remove raw plan model duplicate only**
+
+If `tests/wf_api/test_raw_workflow_plan_extraction.py` covers core step and edge
+model parsing, delete:
+
+```text
+tests/wf_mcp/workflow_surface/test_runs.py::test_raw_workflow_plan_uses_core_step_and_edge_models
+```
+
+Mark the ledger:
+
+```text
+remove | RawWorkflowPlan extraction is canonical in wf_api model tests.
+```
+
+- [ ] **Step 2: Keep run deployment behavior tests**
+
+Keep all remaining tests in `tests/wf_mcp/workflow_surface/test_runs.py`.
+
+Reason: they cover persisted run records, trace slicing with MCP `TraceRange`,
+response Pydantic model validation, logical source binding, schema drift, and
+reducer dependency integration through the handler/service stack.
+
+- [ ] **Step 3: Keep wrapper tests**
+
+Do not delete tests in `tests/wf_mcp/workflow_surface/test_wrappers.py` in this
+pass.
+
+Reason: wrapper direct calls and deployment-bound wrapper calls are meaningful
+handler integration tests even if `WorkflowCapabilityApi` also has lower-level
+coverage.
+
+- [ ] **Step 4: Run run/wrapper tests**
+
+Run:
+
+```bash
+uv run pytest tests/wf_api/test_run_api.py tests/wf_api/test_raw_workflow_plan_extraction.py tests/wf_mcp/workflow_surface/test_runs.py tests/wf_mcp/workflow_surface/test_wrappers.py -q
+```
+
+Expected: pass.
+
+---
+
+## Task 7: Final Review And Verification
+
+**Files:**
+- Modify: `docs/superpowers/plans/2026-06-02-wf-mcp-workflow-surface-test-thinning-ledger.md`
+- Test: all workflow-surface and wf_api tests
+
+- [ ] **Step 1: Ensure ledger has no unclassified rows**
+
+Run:
+
+```bash
+rg -n "unclassified" docs/superpowers/plans/2026-06-02-wf-mcp-workflow-surface-test-thinning-ledger.md
+```
+
+Expected: no matches.
+
+- [ ] **Step 2: Run full focused API and adapter tests**
+
+Run:
+
+```bash
+uv run pytest tests/wf_api tests/wf_mcp/workflow_surface tests/wf_mcp/server/test_tools.py tests/wf_mcp/server/test_config.py -q
+```
+
+Expected: pass.
+
+- [ ] **Step 3: Run lint and type checks**
+
+Run:
+
+```bash
+uv run ruff check tests/wf_api tests/wf_mcp/workflow_surface
+uv run ruff format --check tests/wf_api tests/wf_mcp/workflow_surface
+uv run basedpyright --level error
+```
+
+Expected:
+
+- Ruff commands pass.
+- Basedpyright reports `0 errors`; if it exits nonzero only because workspace
+  enumeration exceeds 10 seconds, report that exact output as an environment
+  issue rather than a type failure.
+
+- [ ] **Step 4: Optional full suite**
+
+Run:
+
+```bash
+uv run pytest -q
+```
+
+Expected: existing suite status remains at least as good as before this pass.
+
+---
+
+## Handoff Report Requirements
+
+When done, report:
+
+- Tests removed, grouped by file.
+- Tests kept intentionally, with reasons for any controversial keeps.
+- Any new or strengthened `wf_api` tests.
+- Ledger path.
+- Verification commands and exact outputs.
+- Deviations from the plan.
+
+## Self-Review
+
+- Spec coverage: the plan preserves “good tests” by requiring a ledger and exact replacement coverage before any deletion.
+- Placeholder scan: no deferred implementation slots; each deletion candidate has a guard and replacement rule.
+- Type consistency: all referenced test paths and test names were taken from the current tree inspection.
@@ -157,6 +157,11 @@ def test_list_capabilities_includes_saved_wrapper() -> None:
    assert row["kind"] == "wrapper_artifact"
    assert row["artifact_id"] == "echo_wrapper"
    assert row["version"] == 1
+    assert row["title"] == "Echo"
+    assert row["description"] == "Reusable echo wrapper."
+    assert row["outcomes"] == ["completed"]
+    assert row["input_fields"] == ["text"]
+    assert row["output_fields"] == ["echoed"]


 def test_inspect_capability_saved_wrapper() -> None:
@@ -177,8 +182,15 @@ def test_inspect_capability_saved_wrapper() -> None:
    assert detail["kind"] == "wrapper_artifact"
    assert detail["artifact_id"] == "echo_wrapper"
    assert detail["outcomes"] == ["completed"]
+    assert "input_schema" in detail
+    assert detail["input_schema"]["properties"]["text"]["type"] == "string"
+    assert detail["output_schema"]["properties"]["echoed"]["type"] == "string"
    hints = detail["wrapper_hints"]
    assert hints["capability_name"] == "workflow.echo_wrapper.v1"
+    assert hints["declared_outcomes"] == ["completed"]
+    assert hints["suggested_wrapper_outcomes"] == ["completed"]
+    assert hints["input_map"] == {"input.text": "text"}
+    assert hints["output_map"] == {"echoed": "state.echoed"}


 def test_call_capability_saved_wrapper() -> None:
@@ -108,6 +108,67 @@ def test_create_draft_workspace_creates_workspace() -> None:

    assert result["workspace_id"] == "echo_ws"
    assert result["revision"] == 1
+    fetched = asyncio.run(
+        api.get_draft_workspace(workspace_id="echo_ws", include_draft=True)
+    )
+    assert fetched["workspace_id"] == "echo_ws"
+    assert fetched["title"] == "Echo Workspace"
+    assert fetched["draft"]["steps"]["echo"]["use"] == "demo.personal.echo_tool"
+
+
+def test_list_draft_workspaces_returns_sorted_summaries_without_drafts() -> None:
+    artifact_store = FileWorkflowArtifactStore(
+        local_temp_root() / "drafts_list_workspaces"
+    )
+    api, _service = _draft_api(artifact_store)
+    asyncio.run(
+        api.create_draft_workspace(
+            workspace_id="b_draft",
+            title="B Draft",
+            draft=_echo_draft(),
+        )
+    )
+    asyncio.run(
+        api.create_draft_workspace(
+            workspace_id="a_draft",
+            title="A Draft",
+            draft=_echo_draft(),
+        )
+    )
+
+    result = asyncio.run(api.list_draft_workspaces())
+
+    assert [workspace["workspace_id"] for workspace in result["workspaces"]] == [
+        "a_draft",
+        "b_draft",
+    ]
+    assert result["workspaces"][0]["title"] == "A Draft"
+    assert "draft" not in result["workspaces"][0]
+
+
+def test_delete_draft_workspace_is_idempotent() -> None:
+    artifact_store = FileWorkflowArtifactStore(
+        local_temp_root() / "drafts_delete_workspace"
+    )
+    api, _service = _draft_api(artifact_store)
+    asyncio.run(
+        api.create_draft_workspace(
+            workspace_id="echo_ws",
+            draft=_echo_draft(),
+        )
+    )
+
+    deleted = asyncio.run(api.delete_draft_workspace(workspace_id="echo_ws"))
+    deleted_again = asyncio.run(api.delete_draft_workspace(workspace_id="echo_ws"))
+    listed = asyncio.run(api.list_draft_workspaces())
+
+    assert deleted["workspace_id"] == "echo_ws"
+    assert deleted["deleted"] is True
+    assert deleted["status"] == "deleted"
+    assert deleted_again["workspace_id"] == "echo_ws"
+    assert deleted_again["deleted"] is False
+    assert deleted_again["status"] == "not_found"
+    assert listed["workspaces"] == []


 def test_patch_draft_workspace_updates_revision() -> None:
@@ -134,6 +195,74 @@ def test_patch_draft_workspace_updates_revision() -> None:
    assert patched["status"] == "valid"


+def test_draft_workspace_patch_helpers_update_revision_and_bindings() -> None:
+    artifact_store = FileWorkflowArtifactStore(
+        local_temp_root() / "drafts_patch_helpers"
+    )
+    api, _service = _draft_api(artifact_store)
+    asyncio.run(
+        api.create_draft_workspace(
+            workspace_id="echo_ws",
+            draft=_echo_draft(),
+        )
+    )
+
+    named = asyncio.run(
+        api.set_draft_name(
+            workspace_id="echo_ws",
+            revision=1,
+            name="echo_v2",
+        )
+    )
+    routed = asyncio.run(
+        api.set_draft_route(
+            workspace_id="echo_ws",
+            revision=2,
+            step_id="echo",
+            outcome="error",
+            target="__end__",
+        )
+    )
+    input_mapped = asyncio.run(
+        api.set_step_input_map(
+            workspace_id="echo_ws",
+            revision=3,
+            step_id="echo",
+            input_map={"input.text": "message"},
+        )
+    )
+    output_mapped = asyncio.run(
+        api.set_step_output_map(
+            workspace_id="echo_ws",
+            revision=4,
+            step_id="echo",
+            output_map={"echoed": "state.echoed"},
+        )
+    )
+    fetched = asyncio.run(
+        api.get_draft_workspace(workspace_id="echo_ws", include_draft=True)
+    )
+
+    assert named["revision"] == 2
+    assert routed["revision"] == 3
+    assert input_mapped["revision"] == 4
+    assert output_mapped["revision"] == 5
+    assert fetched["draft"]["name"] == "echo_v2"
+    assert fetched["draft"]["routes"]["echo"]["error"] == "__end__"
+    assert fetched["draft"]["steps"]["echo"]["input"] == [
+        {
+            "target": {"root": "local", "parts": ["message"]},
+            "path": {"root": "input", "parts": ["text"]},
+        }
+    ]
+    assert fetched["draft"]["steps"]["echo"]["output"] == [
+        {
+            "source": {"root": "local", "parts": ["echoed"]},
+            "target": {"root": "state", "parts": ["echoed"]},
+        }
+    ]
+
+
 def test_validate_draft_workspace_refreshes_status() -> None:
    artifact_store = FileWorkflowArtifactStore(
        local_temp_root() / "drafts_validate_workspace"
@@ -16,56 +16,8 @@ def test_workflow_surface_lists_artifact_catalog_entries() -> None:
    payload = asyncio.run(h.list_artifacts())

    nodes = payload["nodes"]
-    assert len(nodes) == 1
    assert payload["total"] == 1
    assert payload["next_cursor"] is None
    assert nodes[0]["name"] == "workflow.summarize_docs.v1"
    assert nodes[0]["artifact_id"] == "summarize_docs"
-    assert nodes[0]["version"] == 1
-    assert nodes[0]["kind"] == "workflow"
-    assert nodes[0]["required_sources"] == ["context7"]
    assert "plan" not in nodes[0]
-
-
-def test_workflow_surface_pages_and_filters_artifact_catalog_entries() -> None:
-    artifact_store = FileWorkflowArtifactStore(
-        local_temp_root() / "surface_artifact_pages"
-    )
-    artifact_store.save_artifact(artifact())
-    artifact_store.save_artifact(
-        artifact().model_copy(
-            update={
-                "id": "echo_wrapper",
-                "version": 2,
-                "kind": "wrapper",
-                "title": "Echo Wrapper",
-                "description": "Reusable echo wrapper.",
-            }
-        )
-    )
-    artifact_store.save_artifact(
-        artifact().model_copy(
-            update={
-                "id": "browser_click",
-                "title": "Browser Click",
-                "description": "Open a page and wait for a click.",
-            }
-        )
-    )
-    h = handlers(artifact_store)
-
-    first_page = asyncio.run(h.list_artifacts(limit=2))
-    second_page = asyncio.run(
-        h.list_artifacts(cursor=first_page["next_cursor"], limit=2)
-    )
-    wrappers = asyncio.run(h.list_artifacts(kind="wrapper", query="echo"))
-
-    assert first_page["total"] == 3
-    assert first_page["next_cursor"] == "2"
-    assert len(first_page["nodes"]) == 2
-    assert len(second_page["nodes"]) == 1
-    assert second_page["next_cursor"] is None
-    assert wrappers["total"] == 1
-    assert wrappers["nodes"][0]["artifact_id"] == "echo_wrapper"
-    assert wrappers["nodes"][0]["kind"] == "wrapper"
-    assert "plan" not in wrappers["nodes"][0]
@@ -11,7 +11,6 @@ from wf_mcp.workflow_surface import WorkflowSurfaceHandlers
 from ..test_support import echo_tool, local_temp_root
 from .conftest import (
    ContentOnlyOutputAdapter,
-    echo_artifact,
    failing_tool,
    handlers,
 )
@@ -21,30 +20,13 @@ def test_workflow_surface_lists_planner_visible_capabilities() -> None:
    h = handlers(FileWorkflowArtifactStore(local_temp_root() / "surface_caps"))

    payload = asyncio.run(h.list_capabilities(limit=2))
-    names = [capability["name"] for capability in payload["capabilities"]]
    first = payload["capabilities"][0]

-    assert len(names) == 2
+    assert len(payload["capabilities"]) == 2
    assert payload["total"] >= 2
    assert payload["next_cursor"] == "2"
-    assert "description" in first
-    assert "source_id" in first
    assert first["kind"] == "node_spec"
-    assert "input_fields" in first
-    assert "output_fields" in first
    assert "input_schema" not in first
-    assert "wf.admin.list_sources" not in names
-
-
-def test_workflow_surface_filters_stdlib_capabilities_by_source() -> None:
-    h = handlers(FileWorkflowArtifactStore(local_temp_root() / "surface_filtered_caps"))
-
-    payload = asyncio.run(h.list_capabilities(source_id="wf.std", query="truthy"))
-
-    assert [capability["name"] for capability in payload["capabilities"]] == [
-        "wf.std.truthy"
-    ]
-    assert payload["capabilities"][0]["source_id"] == "wf.std"


 def test_workflow_surface_call_capability_returns_structured_error() -> None:
@@ -78,36 +60,6 @@ def test_workflow_surface_call_capability_returns_structured_error() -> None:
    assert "upstream exploded" in payload["diagnostics"][0]["message"]


-def test_workflow_surface_lists_saved_wrapper_capabilities() -> None:
-    artifact_store = FileWorkflowArtifactStore(
-        local_temp_root() / "surface_wrapper_caps"
-    )
-    artifact_store.save_artifact(
-        echo_artifact().model_copy(
-            update={
-                "id": "echo_wrapper",
-                "kind": "wrapper",
-                "description": "Reusable echo wrapper.",
-            }
-        )
-    )
-    artifact_store.save_artifact(echo_artifact())
-    h = handlers(artifact_store)
-
-    payload = asyncio.run(h.list_capabilities(source_id="workflow", query="echo"))
-
-    names = [capability["name"] for capability in payload["capabilities"]]
-    assert names == ["workflow.echo_wrapper.v1"]
-    assert payload["capabilities"][0]["source_id"] == "workflow"
-    assert payload["capabilities"][0]["kind"] == "wrapper_artifact"
-    assert payload["capabilities"][0]["artifact_id"] == "echo_wrapper"
-    assert payload["capabilities"][0]["version"] == 1
-    assert payload["capabilities"][0]["title"] == "Echo"
-    assert payload["capabilities"][0]["outcomes"] == ["completed"]
-    assert payload["capabilities"][0]["input_fields"] == ["text"]
-    assert payload["capabilities"][0]["output_fields"] == ["echoed"]
-
-
 def test_workflow_surface_inspects_one_capability() -> None:
    h = handlers(FileWorkflowArtifactStore(local_temp_root() / "surface_inspect_cap"))

@@ -143,33 +95,6 @@ def test_workflow_surface_inspect_capability_includes_wrapper_hints() -> None:
    assert hints["outcome_policy"] == "preserve_declared"


-def test_workflow_surface_inspects_saved_wrapper_capability() -> None:
-    artifact_store = FileWorkflowArtifactStore(
-        local_temp_root() / "surface_inspect_wrapper_cap"
-    )
-    artifact_store.save_artifact(
-        echo_artifact().model_copy(update={"id": "echo_wrapper", "kind": "wrapper"})
-    )
-    h = handlers(artifact_store)
-
-    payload = asyncio.run(
-        h.inspect_capability(qualified_name="workflow.echo_wrapper.v1")
-    )
-
-    assert payload["name"] == "workflow.echo_wrapper.v1"
-    assert payload["source_id"] == "workflow"
-    assert payload["kind"] == "wrapper_artifact"
-    assert payload["artifact_id"] == "echo_wrapper"
-    assert payload["outcomes"] == ["completed"]
-    assert "input_schema" in payload
-    hints = payload["wrapper_hints"]
-    assert hints["capability_name"] == "workflow.echo_wrapper.v1"
-    assert hints["declared_outcomes"] == ["completed"]
-    assert hints["suggested_wrapper_outcomes"] == ["completed"]
-    assert hints["input_map"] == {"input.text": "text"}
-    assert hints["output_map"] == {"echoed": "state.echoed"}
-
-
 def test_workflow_surface_does_not_auto_map_raw_mcp_content_blocks() -> None:
    artifact_store = FileWorkflowArtifactStore(
        local_temp_root() / "surface_content_only_content_hint"
@@ -20,20 +20,6 @@ from .conftest import (
 )


-def test_workflow_surface_validates_draft_without_saving() -> None:
-    artifact_store = FileWorkflowArtifactStore(
-        local_temp_root() / "surface_draft_validate"
-    )
-    h = handlers(artifact_store)
-
-    payload = asyncio.run(h.validate_draft(draft=echo_draft()))
-
-    assert payload["status"] == "valid"
-    assert payload["diagnostics"] == []
-    assert payload["compiled_plan"]["nodes"][0]["type"] == "node"
-    assert not artifact_store.list_artifacts()
-
-
 def test_workflow_surface_rejects_unknown_draft_route_outcome_when_spec_is_known() -> (
    None
 ):
@@ -129,177 +115,6 @@ def test_workflow_surface_draft_artifact_requires_std_self_binding() -> None:
    assert payload["diagnostics"][0]["logical_ref"] == "wf.std.replace"


-def test_workflow_surface_patches_draft_without_saving() -> None:
-    artifact_store = FileWorkflowArtifactStore(
-        local_temp_root() / "surface_draft_patch"
-    )
-    h = handlers(artifact_store)
-
-    payload = asyncio.run(
-        h.patch_draft(
-            draft=echo_draft(),
-            patch=[
-                {
-                    "op": "replace",
-                    "path": "/steps/echo/input/0/target/parts/0",
-                    "value": "message",
-                }
-            ],
-        )
-    )
-
-    assert payload["status"] == "valid"
-    assert payload["draft"]["steps"]["echo"]["input"][0]["target"] == {
-        "root": "local",
-        "parts": ["message"],
-    }
-    assert not artifact_store.list_artifacts()
-
-
-def test_workflow_surface_creates_and_gets_draft_workspace() -> None:
-    artifact_store = FileWorkflowArtifactStore(local_temp_root() / "surface_workspace")
-    h = handlers(artifact_store)
-
-    created = asyncio.run(
-        h.create_draft_workspace(
-            workspace_id="echo_draft",
-            title="Echo Draft",
-            draft=echo_draft(),
-        )
-    )
-    fetched = asyncio.run(
-        h.get_draft_workspace(
-            workspace_id="echo_draft",
-            include_draft=True,
-        )
-    )
-
-    assert created["workspace_id"] == "echo_draft"
-    assert created["revision"] == 1
-    assert fetched["draft"]["steps"]["echo"]["use"] == "demo.personal.echo_tool"
-
-
-def test_workflow_surface_lists_draft_workspaces() -> None:
-    artifact_store = FileWorkflowArtifactStore(
-        local_temp_root() / "surface_workspace_list"
-    )
-    h = handlers(artifact_store)
-    asyncio.run(
-        h.create_draft_workspace(
-            workspace_id="b_draft",
-            draft=echo_draft(),
-            title="B Draft",
-        )
-    )
-    asyncio.run(
-        h.create_draft_workspace(
-            workspace_id="a_draft",
-            draft=echo_draft(),
-            title="A Draft",
-        )
-    )
-
-    payload = asyncio.run(h.list_draft_workspaces())
-
-    assert [workspace["workspace_id"] for workspace in payload["workspaces"]] == [
-        "a_draft",
-        "b_draft",
-    ]
-    assert payload["workspaces"][0]["title"] == "A Draft"
-    assert "draft" not in payload["workspaces"][0]
-
-
-def test_workflow_surface_deletes_draft_workspace() -> None:
-    artifact_store = FileWorkflowArtifactStore(
-        local_temp_root() / "surface_workspace_delete"
-    )
-    h = handlers(artifact_store)
-    asyncio.run(
-        h.create_draft_workspace(
-            workspace_id="echo_draft",
-            draft=echo_draft(),
-        )
-    )
-
-    deleted = asyncio.run(h.delete_draft_workspace(workspace_id="echo_draft"))
-    deleted_again = asyncio.run(h.delete_draft_workspace(workspace_id="echo_draft"))
-    listed = asyncio.run(h.list_draft_workspaces())
-
-    assert deleted["deleted"] is True
-    assert deleted["status"] == "deleted"
-    assert deleted_again["deleted"] is False
-    assert deleted_again["status"] == "not_found"
-    assert listed["workspaces"] == []
-
-
-def test_workflow_surface_patch_helpers_update_draft_workspace() -> None:
-    artifact_store = FileWorkflowArtifactStore(
-        local_temp_root() / "surface_workspace_patch_helpers"
-    )
-    h = handlers(artifact_store)
-    asyncio.run(
-        h.create_draft_workspace(
-            workspace_id="echo_draft",
-            draft=echo_draft(),
-        )
-    )
-
-    named = asyncio.run(
-        h.set_draft_name(
-            workspace_id="echo_draft",
-            revision=1,
-            name="echo_v2",
-        )
-    )
-    routed = asyncio.run(
-        h.set_draft_route(
-            workspace_id="echo_draft",
-            revision=2,
-            step_id="echo",
-            outcome="error",
-            target="__end__",
-        )
-    )
-    input_mapped = asyncio.run(
-        h.set_step_input_map(
-            workspace_id="echo_draft",
-            revision=3,
-            step_id="echo",
-            input_map={"input.text": "message"},
-        )
-    )
-    output_mapped = asyncio.run(
-        h.set_step_output_map(
-            workspace_id="echo_draft",
-            revision=4,
-            step_id="echo",
-            output_map={"echoed": "state.echoed"},
-        )
-    )
-    fetched = asyncio.run(
-        h.get_draft_workspace(workspace_id="echo_draft", include_draft=True)
-    )
-
-    assert named["revision"] == 2
-    assert routed["revision"] == 3
-    assert input_mapped["revision"] == 4
-    assert output_mapped["revision"] == 5
-    assert fetched["draft"]["name"] == "echo_v2"
-    assert fetched["draft"]["routes"]["echo"]["error"] == "__end__"
-    assert fetched["draft"]["steps"]["echo"]["input"] == [
-        {
-            "target": {"root": "local", "parts": ["message"]},
-            "path": {"root": "input", "parts": ["text"]},
-        }
-    ]
-    assert fetched["draft"]["steps"]["echo"]["output"] == [
-        {
-            "source": {"root": "local", "parts": ["echoed"]},
-            "target": {"root": "state", "parts": ["echoed"]},
-        }
-    ]
-
-
 def test_workflow_surface_validates_draft_workspace_with_live_outcomes() -> None:
    artifact_store = FileWorkflowArtifactStore(
        local_temp_root() / "surface_workspace_validate"
@@ -331,30 +146,6 @@ def test_workflow_surface_validates_draft_workspace_with_live_outcomes() -> None
    assert fetched["status"] == "invalid"


-def test_workflow_surface_patches_draft_workspace_by_revision() -> None:
-    artifact_store = FileWorkflowArtifactStore(
-        local_temp_root() / "surface_workspace_patch"
-    )
-    h = handlers(artifact_store)
-    asyncio.run(
-        h.create_draft_workspace(
-            workspace_id="echo_draft",
-            draft=echo_draft(),
-        )
-    )
-
-    patched = asyncio.run(
-        h.patch_draft_workspace(
-            workspace_id="echo_draft",
-            revision=1,
-            patch=[{"op": "replace", "path": "/name", "value": "echo_v2"}],
-        )
-    )
-
-    assert patched["revision"] == 2
-    assert patched["status"] == "valid"
-
-
 def test_workflow_surface_creates_minimal_draft_workspace_with_error_route() -> None:
    artifact_store = FileWorkflowArtifactStore(
        local_temp_root() / "surface_minimal_workspace"