lda.chat/lda-wf
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
Files
8d714f625c297b1a987d135cb1c9abef2b2cdece
lda-wf/docs/historical/superpowers/plans/2026-07-22-nested-local-draft-bindings.md
T

28 KiB
Raw Blame History

Nested Local Draft Bindings 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: Let focused draft authoring bind nested node-local capability paths while centralizing reusable JSON Schema path lookup and projection.

Architecture: Extend wf_api.schema_projection with one bounded local-reference resolver, one shared path-existence query, and one nested source-to-target projection operation. WorkflowDraftAuthoringApi continues parsing paths with GraphSourcePath and LocalPath, while existing payload serializers, RPC request shapes, and revision-checked patching remain unchanged. CLI parsing validates implied rootless local targets but does not introduce another path model.

Tech Stack: Python 3.14, Pydantic path types, jsonschema Draft 2020-12 validation, Typer CLI, FastAPI JSON-RPC transport, pytest, Ruff, basedpyright.

Global Constraints

  • Preserve revision precedence: after intrinsic request validation, a stale revision returns revision_conflict before workspace-, path-, or capability-derived semantic errors.
  • Use canonical GraphSourcePath and LocalPath; do not add duplicate Pydantic path models or transport endpoint unions.
  • Keep transport request and response shapes unchanged; nested paths remain strings at RPC boundaries.
  • Support only bounded local references #/$defs/... and #/definitions/...; do not build a general JSON Schema resolver or support remote references.
  • Preserve existing single-segment syntax and semantics.
  • Preserve existing public-output lowering through durable state.
  • Do not add literal bindings, fan-out-safe binding lists, structured-input assembly, step metadata, TypeScript RPC parity, or route/revision/runtime changes.
  • Add comments or docstrings around local-reference traversal and other non-obvious schema logic.
  • Run focused tests before broad regression suites; do not run the repository-wide pytest suite for this slice.

Task 1: Centralize JSON Schema Path Lookup And Projection

Files:

  • Modify: src/wf_api/schema_projection.py
  • Modify: src/wf_api/drafts.py
  • Modify: src/wf_api/draft_authoring.py
  • Test: tests/wf_api/test_schema_projection.py

Interfaces:

  • Produces: schema_path_exists(schema: Mapping[str, Any], parts: Sequence[str]) -> bool

  • Produces: project_schema_path_to_schema_path(*, target_schema: JsonObject, source_schema: JsonObject, source_parts: tuple[str, ...], target_parts: tuple[str, ...], allow_existing_equivalent: bool = False) -> JsonObject

  • Preserves: project_property_to_schema_path(...) as a root-property compatibility wrapper.

  • Removes: private duplicate _schema_path_exists implementations from wf_api.drafts and wf_api.draft_authoring.

  • Step 1: Add failing tests for inline and referenced nested source paths

    Extend tests/wf_api/test_schema_projection.py imports with the two new public operations, then add these cases:

    def test_project_schema_path_copies_inline_nested_source() -> None:
    projected = project_schema_path_to_schema_path(
        target_schema={"type": "object", "properties": {}},
        source_schema={
            "type": "object",
            "properties": {
                "report": {
                    "type": "object",
                    "properties": {"title": {"type": "string"}},
                }
            },
        },
        source_parts=("report", "title"),
        target_parts=("document", "title"),
    )

    assert projected["properties"]["document"]["properties"]["title"] == {
        "type": "string"
    }


def test_project_schema_path_traverses_pydantic_defs_reference() -> None:
    projected = project_schema_path_to_schema_path(
        target_schema={"type": "object", "properties": {}},
        source_schema={
            "type": "object",
            "properties": {"report": {"$ref": "#/$defs/Report"}},
            "$defs": {
                "Report": {
                    "type": "object",
                    "properties": {
                        "markdown": {"$ref": "#/$defs/Markdown"}
                    },
                },
                "Markdown": {"type": "string", "minLength": 1},
            },
        },
        source_parts=("report", "markdown"),
        target_parts=("report", "markdown"),
    )

    assert projected["properties"]["report"]["properties"]["markdown"] == {
        "$ref": "#/$defs/Markdown"
    }
    assert projected["$defs"]["Markdown"]["minLength"] == 1


def test_project_schema_path_traverses_legacy_definitions_reference() -> None:
    projected = project_schema_path_to_schema_path(
        target_schema={"type": "object", "properties": {}},
        source_schema={
            "type": "object",
            "properties": {"report": {"$ref": "#/definitions/Report"}},
            "definitions": {
                "Report": {
                    "type": "object",
                    "properties": {"title": {"type": "string"}},
                }
            },
        },
        source_parts=("report", "title"),
        target_parts=("title",),
    )

    assert projected["properties"]["title"] == {"type": "string"}
    assert "Report" in projected["definitions"]

  • Step 2: Add failing tests for existence and precise failures

    Add tests that pin the bounded behavior:

    def test_schema_path_exists_follows_local_defs() -> None:
    schema = {
        "type": "object",
        "properties": {"report": {"$ref": "#/$defs/Report"}},
        "$defs": {
            "Report": {
                "type": "object",
                "properties": {"title": {"type": "string"}},
            }
        },
    }

    assert schema_path_exists(schema, ("report", "title")) is True
    assert schema_path_exists(schema, ("report", "missing")) is False


def test_project_schema_path_rejects_missing_nested_source() -> None:
    with pytest.raises(ValueError, match="source schema path 'report.missing' is not declared"):
        project_schema_path_to_schema_path(
            target_schema={"type": "object", "properties": {}},
            source_schema={
                "type": "object",
                "properties": {
                    "report": {"type": "object", "properties": {}}
                },
            },
            source_parts=("report", "missing"),
            target_parts=("value",),
        )


def test_project_schema_path_rejects_scalar_source_ancestor() -> None:
    with pytest.raises(ValueError, match="source schema path 'report' is not an object"):
        project_schema_path_to_schema_path(
            target_schema={"type": "object", "properties": {}},
            source_schema={
                "type": "object",
                "properties": {"report": {"type": "string"}},
            },
            source_parts=("report", "title"),
            target_parts=("title",),
        )


def test_project_schema_path_rejects_remote_reference() -> None:
    with pytest.raises(ValueError, match="unsupported reference 'https://example.com/report.json'"):
        project_schema_path_to_schema_path(
            target_schema={"type": "object", "properties": {}},
            source_schema={
                "type": "object",
                "properties": {
                    "report": {"$ref": "https://example.com/report.json"}
                },
            },
            source_parts=("report", "title"),
            target_parts=("title",),
        )

    Retain the existing target-conflict and equivalent-target tests; they are regression coverage for the generalized operation.

  • Step 3: Run the schema tests and confirm the new imports fail

    Run:

    uv run pytest tests/wf_api/test_schema_projection.py -q -n 0

    Expected: collection fails because schema_path_exists and project_schema_path_to_schema_path are not exported yet.

  • Step 4: Implement one bounded path resolver and the generalized projector

    In src/wf_api/schema_projection.py:

    • import Mapping and Sequence from collections.abc;
    • add a private _resolve_local_reference(root_schema, candidate, label) helper that accepts only #/$defs/<name> and #/definitions/<name>, checks the selected definition is a mapping, and raises a precise ValueError for unsupported or unresolved references;
    • add a private _schema_at_path(root_schema, parts, *, label) helper that resolves local references before reading each intermediate object's properties;
    • make schema_path_exists call _schema_at_path and return False for missing paths, non-object ancestors, or unresolved references;
    • move the current target-path insertion and definition-block merge into project_schema_path_to_schema_path after selecting the nested source schema;
    • make project_property_to_schema_path delegate with source_parts=(source_field,) and translate root-field errors where needed to preserve existing wording.

    The implementation must retain the selected leaf unchanged. For example, when the leaf is {"$ref": "#/$defs/Markdown"}, copy that reference and merge the source definition blocks rather than replacing the leaf with the resolved definition.

  • Step 5: Replace both duplicate existence helpers with the shared function

    In src/wf_api/drafts.py and src/wf_api/draft_authoring.py:

    from .schema_projection import schema_path_exists

    Delete each private _schema_path_exists definition and replace its callers with schema_path_exists(...). Remove now-unused Mapping or Sequence imports only when no other symbol in that module needs them.

  • Step 6: Run focused tests and quality checks

    Run:

    uv run pytest tests/wf_api/test_schema_projection.py tests/wf_api/test_drafts_service.py -q -n 0
uv run ruff check src/wf_api/schema_projection.py src/wf_api/drafts.py src/wf_api/draft_authoring.py tests/wf_api/test_schema_projection.py
uv run ruff format --check src/wf_api/schema_projection.py src/wf_api/drafts.py src/wf_api/draft_authoring.py tests/wf_api/test_schema_projection.py
uv run basedpyright --level error src/wf_api/schema_projection.py src/wf_api/drafts.py src/wf_api/draft_authoring.py

    Expected: all schema projection and draft service tests pass; all quality checks are clean.

  • Step 7: Commit the shared schema operation

    git add src/wf_api/schema_projection.py src/wf_api/drafts.py src/wf_api/draft_authoring.py tests/wf_api/test_schema_projection.py
git commit -m "refactor: centralize schema path projection"

Task 2: Support Nested Local Paths In Focused Draft Bind

Files:

  • Modify: src/wf_api/draft_authoring.py
  • Test: tests/wf_api/test_drafts_service.py

Interfaces:

  • Consumes: project_schema_path_to_schema_path(...) and schema_path_exists(...) from Task 1.

  • Preserves: WorkflowDraftAuthoringApi.bind_draft(...) request and response signatures.

  • Produces: complete nested LocalPath values in existing input/output binding payloads.

  • Step 1: Add a nested capability fixture with Pydantic-style definitions

    Add this test-only capability beside the existing echo/snapshot fixtures in tests/wf_api/test_drafts_service.py:

    class _ReportInputValue(BaseModel):
    title: str


class _NestedReportInput(BaseModel):
    report: _ReportInputValue


class _ReportOutputValue(BaseModel):
    markdown: str


class _NestedReportOutput(BaseModel):
    report: _ReportOutputValue


@node(name="nested_report", outcomes=("ok",))
def _nested_report(payload: _NestedReportInput) -> _NestedReportOutput:
    return _NestedReportOutput(
        report=_ReportOutputValue(markdown=f"# {payload.report.title}")
    )

    Register it in each new test with the same existing connection setup used for _snapshot_tool:

    service.register_specs("demo.personal", _nested_report)

  • Step 2: Add failing input/state-to-local tests

    Add tests that call:

    await authoring.bind_draft(
    workspace_id="nested",
    revision=1,
    step_id="render",
    source_path="input.title",
    target_path="local.report.title",
)

    and, in a separate workspace:

    await authoring.bind_draft(
    workspace_id="nested_state",
    revision=1,
    step_id="render",
    source_path="state.report.title",
    target_path="local.report.title",
)

    Assert field-level results:

    assert draft["input_schema"]["properties"]["title"]["type"] == "string"
assert draft["steps"]["render"]["input"] == [
    {"target": "report.title", "path": "input.title"}
]

    For the state case, assert the nested state schema remains valid and the stored target is exactly report.title.

  • Step 3: Add failing nested local-output tests

    Cover both supported directions:

    local.report.markdown -> state.report.markdown
local.report.markdown -> output.report.markdown

    For state, assert:

    assert draft["steps"]["render"]["output"] == [
    {"source": "report.markdown", "target": "state.report.markdown"}
]
assert draft["state_schema"]["properties"]["report"]["properties"]["markdown"]["type"] == "string"

    For public output, assert the existing lowering contract:

    assert draft["steps"]["render"]["output"] == [
    {"source": "report.markdown", "target": "state.report.markdown"}
]
assert draft["output"] == [
    {"path": "state.report.markdown", "target": "report.markdown"}
]

    Also assert both state_schema and output_schema contain the projected nested string schema.

  • Step 4: Add failure atomicity and revision-precedence tests

    Add one current-revision test using local.report.missing -> state.report.missing. Assert the complete path appears in the ValueError, then fetch the workspace and assert revision == 1 and the step output remains unchanged.

    Add one stale-revision test with the same invalid nested path after first moving the workspace to revision 2. Assert the returned diagnostic code is revision_conflict and no nested-path ValueError escapes.

  • Step 5: Run the new API tests and confirm the one-field restriction fails

    Run:

    uv run pytest tests/wf_api/test_drafts_service.py -q -n 0 -k "nested and bind"

    Expected: nested local endpoints fail with local path must name one capability field.

  • Step 6: Replace root-field handling with complete local parts

    In WorkflowDraftAuthoringApi.bind_draft:

    • delete _local_field;
    • use the already-parsed source_parts or target_parts tuple for local endpoints;
    • call project_schema_path_to_schema_path with the complete local path as source_parts;
    • serialize complete local paths with format_toml_path_segments(local_parts) when updating the focused input/output maps;
    • retain schema_path_exists reuse for already-declared workflow input/state paths;
    • retain allow_existing_equivalent=True for state/output projections;
    • retain existing public-output replacement and state-lowering behavior.

    The relevant input branch becomes structurally equivalent to:

    local_path = format_toml_path_segments(target_parts)
projected = (
    target_schema
    if schema_path_exists(target_schema, source_parts)
    else project_schema_path_to_schema_path(
        target_schema=target_schema,
        source_schema=input_schema,
        source_parts=target_parts,
        target_parts=source_parts,
    )
)
input_map = {
    **_input_map_from_payload(step.get("input", [])),
    source_path: local_path,
}

  • Step 7: Run focused API and canonical-model regressions

    Run:

    uv run pytest tests/wf_api/test_drafts_service.py tests/core/test_nested_mappings.py tests/core/test_mapping_validation.py -q -n 0
uv run ruff check src/wf_api/draft_authoring.py tests/wf_api/test_drafts_service.py
uv run ruff format --check src/wf_api/draft_authoring.py tests/wf_api/test_drafts_service.py
uv run basedpyright --level error src/wf_api/draft_authoring.py

    Expected: nested and existing single-field cases pass, and canonical nested mapping validation remains green.

  • Step 8: Commit focused nested bind support

    git add src/wf_api/draft_authoring.py tests/wf_api/test_drafts_service.py
git commit -m "feat: bind nested local draft paths"

Task 3: Support Nested Capability Maps Through CLI And RPC

Files:

  • Modify: src/wf_api/draft_authoring.py
  • Modify: src/wf_cli/commands/draft_options.py
  • Modify: src/wf_cli/commands/drafts.py
  • Modify: src/wf_cli/commands/draft_add.py
  • Modify: src/wf_mcp/workflow_surface/models.py
  • Test: tests/wf_api/test_drafts_service.py
  • Test: tests/wf_transport_rpc_http/test_app.py
  • Test: tests/wf_transport_rpc_http/test_client.py
  • Test: tests/wf_cli/test_app.py
  • Test: tests/wf_cli/test_remote_target.py

Interfaces:

  • Consumes: complete nested schema selection from Task 1.

  • Preserves: add_step_from_capability(..., input_map: dict[str, str] | None, ...) and all transport envelopes.

  • Produces: validated rootless local map targets such as report.title in CLI commands.

  • Step 1: Add failing capability-add API coverage

    In tests/wf_api/test_drafts_service.py, add a test that creates an empty-schema draft and calls:

    result = await authoring.add_step_from_capability(
    workspace_id="nested_add",
    revision=1,
    step_id="render",
    capability_name="demo.personal.nested_report",
    routes={"ok": "__end__"},
    input_map={"input.title": "report.title"},
    bind_outputs={},
)

    Assert revision 2, stored target report.title, and projected workflow input schema title: string. Validate or compile the resulting draft through the existing draft validation API and assert status == "valid".

  • Step 2: Add failing CLI parser and help tests

    In tests/wf_cli/test_app.py, add assertions that:

    • wf draft set-input ... --map input.title=report.title reaches the handler with the nested string unchanged;
    • wf draft add capability ... --input input.title=report.title reaches the handler unchanged;
    • --map input.title=local.report.title exits with code 2 and suggests input.title=report.title;
    • malformed rootless local paths fail before RPC;
    • wf draft bind --help describes explicit rooted endpoints such as local.report.title;
    • set-input and capability-add help describe rootless local paths such as report.title.

  • Step 3: Add failing transport preservation tests

    In tests/wf_transport_rpc_http/test_app.py and tests/wf_transport_rpc_http/test_client.py, extend the existing bind and add-capability tests with nested values. Assert request delegation and stored results preserve:

    source_path == "input.title"
target_path == "local.report.title"
input_map == {"input.title": "report.title"}

    In tests/wf_cli/test_remote_target.py, assert the remote CLI emits the existing JSON-RPC method names and unchanged nested strings. Do not add new RPC models or methods.

  • Step 4: Run focused tests and confirm nested projection/help failures

    Run:

    uv run pytest tests/wf_api/test_drafts_service.py tests/wf_transport_rpc_http/test_app.py tests/wf_transport_rpc_http/test_client.py tests/wf_cli/test_app.py tests/wf_cli/test_remote_target.py -q -n 0 -k "nested or bind_help or add_step_from_capability"

    Expected: the add-capability test fails because nested local input paths are skipped; CLI validation/help assertions fail until implemented.

  • Step 5: Project nested capability input paths instead of skipping them

    In WorkflowDraftAuthoringApi.add_step_from_capability:

    • keep parsing graph sources with _graph_parts and targets with LocalPath.parse;
    • remove len(local_parts) != 1 from the skip condition;
    • call project_schema_path_to_schema_path with source_parts=local_parts;
    • continue reusing existing workflow source paths with schema_path_exists;
    • preserve the current revision preflight and atomic patch.

    Invalid canonical paths may retain current request-layer behavior, but a valid nested local path must never silently skip schema projection.

  • Step 6: Validate rootless local targets and clarify help

    In _parse_step_input_map_flags, retain the rooted-target repair first, then validate each rootless target:

    try:
    LocalPath.parse(target)
except PathResolutionError as exc:
    raise typer.BadParameter(
        f"{option_name} target {target!r} must be a rootless node-local "
        "path such as report.title or ."
    ) from exc

    Update command docstrings/help in drafts.py and draft_add.py to distinguish:

    bind endpoint: --from input.title --to local.report.title
implied local map: --map input.title=report.title

    Update only descriptions in BindDraftRequest and the capability-add request model in src/wf_mcp/workflow_surface/models.py; do not alter fields or add validation models.

  • Step 7: Run API, RPC, client, and CLI tests

    Run:

    uv run pytest tests/wf_api/test_drafts_service.py tests/wf_transport_rpc_http/test_app.py tests/wf_transport_rpc_http/test_client.py tests/wf_cli/test_app.py tests/wf_cli/test_remote_target.py -q -n 0
uv run ruff check src/wf_api/draft_authoring.py src/wf_cli/commands/draft_options.py src/wf_cli/commands/drafts.py src/wf_cli/commands/draft_add.py src/wf_mcp/workflow_surface/models.py tests/wf_api/test_drafts_service.py tests/wf_transport_rpc_http/test_app.py tests/wf_transport_rpc_http/test_client.py tests/wf_cli/test_app.py tests/wf_cli/test_remote_target.py
uv run ruff format --check src/wf_api/draft_authoring.py src/wf_cli/commands/draft_options.py src/wf_cli/commands/drafts.py src/wf_cli/commands/draft_add.py src/wf_mcp/workflow_surface/models.py tests/wf_api/test_drafts_service.py tests/wf_transport_rpc_http/test_app.py tests/wf_transport_rpc_http/test_client.py tests/wf_cli/test_app.py tests/wf_cli/test_remote_target.py
uv run basedpyright --level error src/wf_api/draft_authoring.py src/wf_cli src/wf_mcp/workflow_surface/models.py

    Expected: all focused layers pass with no request-schema changes.

  • Step 8: Commit nested capability-map and surface support

    git add src/wf_api/draft_authoring.py src/wf_cli/commands/draft_options.py src/wf_cli/commands/drafts.py src/wf_cli/commands/draft_add.py src/wf_mcp/workflow_surface/models.py tests/wf_api/test_drafts_service.py tests/wf_transport_rpc_http/test_app.py tests/wf_transport_rpc_http/test_client.py tests/wf_cli/test_app.py tests/wf_cli/test_remote_target.py
git commit -m "feat: accept nested local capability maps"

Task 4: Update User Guidance, Close Covered Issues, And Verify The Slice

Files:

  • Modify: skills/wf-cli/SKILL.md
  • Modify: skills/wf-workflow/references/draft-workspaces.md
  • Modify: docs/wf_cli.md
  • Modify: ISSUES.md
  • Modify: docs/current_roadmap.md
  • Move after verification: docs/superpowers/plans/2026-07-22-nested-local-draft-bindings.md to docs/historical/superpowers/plans/2026-07-22-nested-local-draft-bindings.md

Interfaces:

  • Documents: explicit rooted bind endpoints versus implied rootless map targets.

  • Closes only: nested bind rejection, capability-add nested projection skip, and stale CLI/agent guidance.

  • Leaves open: structured-input assembly, literals, fan-out maps, nested workflow-output source projection, metadata/update-step, and TypeScript parity.

  • Step 1: Update CLI and agent-facing examples

    Replace “bare field” wording with “rootless node-local path” in all three live user-facing documents. Include both forms together:

    wf draft bind report_ws --revision 2 --step render \
  --from input.title --to local.report.title

wf draft set-input report_ws --revision 3 --step render \
  --map input.title=report.title

    Explain that wf draft bind names both rooted endpoints, while set-input and capability-add already imply the local side and therefore accept rootless targets. Do not claim support for literals, fan-out, or atomic object assembly.

  • Step 2: Update issue and roadmap state precisely

    In ISSUES.md, check only these three entries:

    • nested wf draft bind local paths;
    • nested capability-step schema projection;
    • CLI help and agent instruction coverage.

    Keep every other Draft data-shaping parity entry unchecked.

    In docs/current_roadmap.md, add a completed entry summarizing nested local bind/capability map parity and link to:

    historical/superpowers/plans/2026-07-22-nested-local-draft-bindings.md

  • Step 3: Run the complete focused verification matrix

    Run:

    uv run pytest tests/wf_api/test_schema_projection.py tests/wf_api/test_drafts_service.py tests/wf_transport_rpc_http/test_app.py tests/wf_transport_rpc_http/test_client.py tests/wf_cli/test_app.py tests/wf_cli/test_remote_target.py tests/core/test_nested_mappings.py tests/core/test_mapping_validation.py -q -n 0
uv run ruff check src/wf_api/schema_projection.py src/wf_api/drafts.py src/wf_api/draft_authoring.py src/wf_cli/commands/draft_options.py src/wf_cli/commands/drafts.py src/wf_cli/commands/draft_add.py src/wf_mcp/workflow_surface/models.py tests/wf_api/test_schema_projection.py tests/wf_api/test_drafts_service.py tests/wf_transport_rpc_http/test_app.py tests/wf_transport_rpc_http/test_client.py tests/wf_cli/test_app.py tests/wf_cli/test_remote_target.py
uv run ruff format --check src/wf_api/schema_projection.py src/wf_api/drafts.py src/wf_api/draft_authoring.py src/wf_cli/commands/draft_options.py src/wf_cli/commands/drafts.py src/wf_cli/commands/draft_add.py src/wf_mcp/workflow_surface/models.py tests/wf_api/test_schema_projection.py tests/wf_api/test_drafts_service.py tests/wf_transport_rpc_http/test_app.py tests/wf_transport_rpc_http/test_client.py tests/wf_cli/test_app.py tests/wf_cli/test_remote_target.py
uv run basedpyright --level error src/wf_api/schema_projection.py src/wf_api/drafts.py src/wf_api/draft_authoring.py src/wf_cli src/wf_mcp/workflow_surface/models.py
git diff --check

    Expected: all focused tests and static checks pass. The pytest count may increase as tests are added; report exact totals rather than copying a planned count.

  • Step 4: Run an independent two-axis review

    Use the requesting-code-review skill against the implementation start commit. Review:

    1. Standards: repository conventions, duplicate schema logic, compatibility-only branches, test quality, comments around reference resolution.
    2. Spec: every requirement in docs/superpowers/specs/2026-07-22-nested-local-draft-bindings-design.md, especially stale-revision precedence and unchanged transport shapes.

    Fix Critical and Important findings, rerun affected focused tests, and record any intentional Minor deferrals in the final report.

  • Step 5: Archive the completed plan and commit documentation

    After all checks pass:

    git mv docs/superpowers/plans/2026-07-22-nested-local-draft-bindings.md docs/historical/superpowers/plans/2026-07-22-nested-local-draft-bindings.md
git add ISSUES.md docs/current_roadmap.md docs/wf_cli.md skills/wf-cli/SKILL.md skills/wf-workflow/references/draft-workspaces.md docs/historical/superpowers/plans/2026-07-22-nested-local-draft-bindings.md
git commit -m "docs: complete nested local draft bindings"

    Confirm git status --short is clean and report changed behavior, exact verification results, deviations, bugs fixed, and review findings.

Self-Review

  • Spec coverage: Task 1 covers shared schema lookup, bounded local refs, definition propagation, validation, conflicts, and duplicate-helper removal. Task 2 covers focused bind in every scoped direction, lowering, atomic failure, and revision precedence. Task 3 covers capability insertion, CLI validation/help, and unchanged RPC delegation. Task 4 covers live docs, issue state, roadmap, archive, and the full verification gate.
  • Placeholder scan: Every code change names the concrete behavior, signature, error, or assertion required. No deferred fill-in steps or generic error-handling instructions remain.
  • Type consistency: Every later task consumes the exact public signatures defined in Task 1. bind_draft, add_step_from_capability, transport request fields, and focused payload serializer signatures remain unchanged.
Reference in New Issue View Git Blame Copy Permalink