agent-lifecycle-kit 2.25.0__tar.gz → 2.26.0__tar.gz

{agent_lifecycle_kit-2.25.0/src/agent_lifecycle_kit.egg-info → agent_lifecycle_kit-2.26.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-lifecycle-kit
-Version: 2.25.0
+Version: 2.26.0
 Summary: Provider-neutral lifecycle control layer that helps coding agents plan, execute, validate, and prove software tasks to completion.
 Author: Agent Lifecycle Kit contributors
 License: Apache-2.0
@@ -41,7 +41,7 @@ ALK provides a verifiable finish for external-agent work with proportionate cont
 ## Quick start
 Start with [install ALK and make the first run](docs/guides/install-and-first-run.md). Release 2.23 adds [compact S0/S1 authoring](docs/reference/tier-proportional-plan-packages.md) with explicit checklist completion and unchanged approval gates. Release 2.22 adds [conservative plan-review reuse](docs/reference/conservative-delta-review.md) with independent current applicability and no automatic full reruns. Release 2.21 adds [command accounting, bounded preparation and response recovery](docs/reference/source-native-workflow-accounting.md). Release 2.20 adds [target-bound initialization](docs/reference/cli.md#release-220-composed-initialization-and-digest-display), [bounded local continuation](docs/reference/workflow-continuation.md) and batched Git snapshots. Release 2.19 adds [safe artifact authoring and planning preflight](docs/reference/safe-artifact-authoring.md), [bounded review handoff](docs/reference/review-mesh.md#bounded-review-handoff) and thin skills with complete references. Release 2.18 bounds process/Git diagnostics, hardens signing and distinguishes OpenCode 1.x/2.x compatibility ([adapter notes](docs/adapters/opencode.md)); 2.17 strengthens [workflow validation and recovery](docs/reference/workflow-completeness.md). Release 2.16 introduced cumulative accounting within verified host capabilities; see [retention maintenance](docs/operations/retention-maintenance.md).
 
-It includes macOS, Linux and Windows instructions, Python 3.11-3.14, [PyPI](https://pypi.org/project/agent-lifecycle-kit/) installation with `python -m pip install agent-lifecycle-kit==2.25.0`, and common `agent-lifecycle version` errors. See [2.15 migration notes](docs/reference/cli.md#migrating-authority-artifacts-to-215) and [authority boundaries](docs/reference/authority-artifact-boundaries.md).
+It includes macOS, Linux and Windows instructions, Python 3.11-3.14, [PyPI](https://pypi.org/project/agent-lifecycle-kit/) installation with `python -m pip install agent-lifecycle-kit==2.26.0`, and common `agent-lifecycle version` errors. See [2.15 migration notes](docs/reference/cli.md#migrating-authority-artifacts-to-215) and [authority boundaries](docs/reference/authority-artifact-boundaries.md).
 
 For a quick start, run:
 
@@ -194,9 +194,9 @@ Adapter installation and support-level details live in [Adapter install](docs/ad
 - Release assets: [Task templates](templates/tasks/README.md), [Neutrality scanning](docs/reference/neutrality.md), and [Release security](docs/security/release-security.md).
 Apache-2.0. See [LICENSE](LICENSE).
 
-## Release 2.25: focused security and maintainability
+## Release 2.26: bounded discovery and precise quality gates
 
-Release 2.25 hardens sensitive imports and launcher checks, repairs named ownership seams, and checks untyped Python function bodies with mypy. See the [current security
-controls](docs/security/release-security.md) and [architecture map](docs/architecture/system-architecture.md). The [staged reading
-pilot](docs/guides/controlled-technical-english.md) preserves full references and approval gates; byte/word measurements do not prove token savings. Existing OpenCode 1.x/2.x
-qualification and unsupported host routes remain unchanged. Local preparation, exact-candidate CI, native final proof and public publication are separate gates.
+Markdown imports bound examined entries, directories, depth and cooperative elapsed time; see the [security control map](docs/security/release-security.md). Tracked-source checks
+reuse the existing bare-token families. [Quality controls](docs/architecture/system-architecture.md)
+require both combined and statement coverage of 76 percent and retain legacy policy interpretation. The existing [reading pilot](docs/guides/controlled-technical-english.md) and host limits
+remain; measurements do not establish token savings or host enforcement.

{agent_lifecycle_kit-2.25.0 → agent_lifecycle_kit-2.26.0}/README.md

@@ -19,7 +19,7 @@ ALK provides a verifiable finish for external-agent work with proportionate cont
 ## Quick start
 Start with [install ALK and make the first run](docs/guides/install-and-first-run.md). Release 2.23 adds [compact S0/S1 authoring](docs/reference/tier-proportional-plan-packages.md) with explicit checklist completion and unchanged approval gates. Release 2.22 adds [conservative plan-review reuse](docs/reference/conservative-delta-review.md) with independent current applicability and no automatic full reruns. Release 2.21 adds [command accounting, bounded preparation and response recovery](docs/reference/source-native-workflow-accounting.md). Release 2.20 adds [target-bound initialization](docs/reference/cli.md#release-220-composed-initialization-and-digest-display), [bounded local continuation](docs/reference/workflow-continuation.md) and batched Git snapshots. Release 2.19 adds [safe artifact authoring and planning preflight](docs/reference/safe-artifact-authoring.md), [bounded review handoff](docs/reference/review-mesh.md#bounded-review-handoff) and thin skills with complete references. Release 2.18 bounds process/Git diagnostics, hardens signing and distinguishes OpenCode 1.x/2.x compatibility ([adapter notes](docs/adapters/opencode.md)); 2.17 strengthens [workflow validation and recovery](docs/reference/workflow-completeness.md). Release 2.16 introduced cumulative accounting within verified host capabilities; see [retention maintenance](docs/operations/retention-maintenance.md).
 
-It includes macOS, Linux and Windows instructions, Python 3.11-3.14, [PyPI](https://pypi.org/project/agent-lifecycle-kit/) installation with `python -m pip install agent-lifecycle-kit==2.25.0`, and common `agent-lifecycle version` errors. See [2.15 migration notes](docs/reference/cli.md#migrating-authority-artifacts-to-215) and [authority boundaries](docs/reference/authority-artifact-boundaries.md).
+It includes macOS, Linux and Windows instructions, Python 3.11-3.14, [PyPI](https://pypi.org/project/agent-lifecycle-kit/) installation with `python -m pip install agent-lifecycle-kit==2.26.0`, and common `agent-lifecycle version` errors. See [2.15 migration notes](docs/reference/cli.md#migrating-authority-artifacts-to-215) and [authority boundaries](docs/reference/authority-artifact-boundaries.md).
 
 For a quick start, run:
 
@@ -172,9 +172,9 @@ Adapter installation and support-level details live in [Adapter install](docs/ad
 - Release assets: [Task templates](templates/tasks/README.md), [Neutrality scanning](docs/reference/neutrality.md), and [Release security](docs/security/release-security.md).
 Apache-2.0. See [LICENSE](LICENSE).
 
-## Release 2.25: focused security and maintainability
+## Release 2.26: bounded discovery and precise quality gates
 
-Release 2.25 hardens sensitive imports and launcher checks, repairs named ownership seams, and checks untyped Python function bodies with mypy. See the [current security
-controls](docs/security/release-security.md) and [architecture map](docs/architecture/system-architecture.md). The [staged reading
-pilot](docs/guides/controlled-technical-english.md) preserves full references and approval gates; byte/word measurements do not prove token savings. Existing OpenCode 1.x/2.x
-qualification and unsupported host routes remain unchanged. Local preparation, exact-candidate CI, native final proof and public publication are separate gates.
+Markdown imports bound examined entries, directories, depth and cooperative elapsed time; see the [security control map](docs/security/release-security.md). Tracked-source checks
+reuse the existing bare-token families. [Quality controls](docs/architecture/system-architecture.md)
+require both combined and statement coverage of 76 percent and retain legacy policy interpretation. The existing [reading pilot](docs/guides/controlled-technical-english.md) and host limits
+remain; measurements do not establish token savings or host enforcement.

{agent_lifecycle_kit-2.25.0 → agent_lifecycle_kit-2.26.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "setuptools.build_meta"
 
 [project]
 name = "agent-lifecycle-kit"
-version = "2.25.0"
+version = "2.26.0"
 description = "Provider-neutral lifecycle control layer that helps coding agents plan, execute, validate, and prove software tasks to completion."
 readme = "README.md"
 requires-python = ">=3.11,<3.15"

{agent_lifecycle_kit-2.25.0 → agent_lifecycle_kit-2.26.0}/src/agent_lifecycle/_version.py

@@ -1,3 +1,3 @@
 """Package version."""
 
-__version__ = "2.25.0"
+__version__ = "2.26.0"

{agent_lifecycle_kit-2.25.0 → agent_lifecycle_kit-2.26.0}/src/agent_lifecycle/contracts/redaction.py

@@ -152,6 +152,12 @@ def contains_local_absolute_path(value: str) -> bool:
     return bool(_LOCAL_PATH.search(protected))
 
 
+def contains_bare_secret_token(value: str) -> bool:
+    """Classify the existing bare-token forms without rewriting text or applying assignment rules."""
+
+    return bool(_BARE_TOKEN.search(value))
+
+
 def redact_value(value: Any) -> tuple[Any, bool]:
     """Redact nested receipt values while preserving JSON-compatible structure."""
 

agent_lifecycle_kit-2.26.0/src/agent_lifecycle/imports/multi_markdown.py

@@ -0,0 +1,238 @@
+"""Deterministic Markdown file and folder import helpers."""
+
+from __future__ import annotations
+
+import os
+import stat
+import time
+from collections.abc import Callable
+from fnmatch import fnmatch
+from functools import partial
+from pathlib import Path
+from typing import Any, TypeVar
+
+from agent_lifecycle.contracts import LifecycleError, canonical_digest, sha256_hex
+from agent_lifecycle.contracts.authority_io import hold_authority_directory, open_authority_read
+from agent_lifecycle.imports.planning import (
+    DEFAULT_MAX_INPUT_BYTES,
+    DEFAULT_TARGET_TOKENS,
+    import_planning_text,
+    require_no_import_secrets,
+)
+
+MARKDOWN_COLLECTION_SCHEMA = "agent-markdown-source-collection.v1"
+DEFAULT_MAX_MARKDOWN_FILES = 32
+_MAX_DISCOVERY_ENTRIES = 4096
+_MAX_DISCOVERY_DIRECTORIES = 256
+_MAX_DISCOVERY_DEPTH = 32
+_DISCOVERY_SECONDS = 5.0
+_T = TypeVar("_T")
+
+
+def import_markdown_collection(
+    source_path: Path,
+    *,
+    package_id: str = "markdown-import",
+    max_input_bytes: int = DEFAULT_MAX_INPUT_BYTES,
+    target_tokens: int = DEFAULT_TARGET_TOKENS,
+    max_files: int = DEFAULT_MAX_MARKDOWN_FILES,
+    dialect_profile: dict[str, Any] | None = None,
+) -> dict[str, Any]:
+    """Import one Markdown file or a directory of Markdown files as draft input."""
+
+    collection = collect_markdown_collection(source_path, max_input_bytes=max_input_bytes, max_files=max_files)
+    text = collection.pop("_combinedText", "")
+    result = import_planning_text(
+        text,
+        source_label=collection["sourceLabel"],
+        package_id=package_id,
+        max_input_bytes=max_input_bytes,
+        target_tokens=target_tokens,
+        dialect_profile=dialect_profile,
+    )
+    result["markdownCollection"] = collection
+    if collection["blockers"]:
+        result["blockers"].extend(collection["blockers"])
+        result["status"] = "FAIL"
+        result["candidateLifecycleStatus"] = "BLOCKED"
+        result["candidatePlan"] = None
+    result["importDigest"] = canonical_digest({key: value for key, value in result.items() if key != "importDigest"})
+    return result
+
+
+def collect_markdown_collection(
+    source_path: Path,
+    *,
+    max_input_bytes: int = DEFAULT_MAX_INPUT_BYTES,
+    max_files: int = DEFAULT_MAX_MARKDOWN_FILES,
+) -> dict[str, Any]:
+    for field, cap in (("maxFiles", max_files), ("maxInputBytes", max_input_bytes)):
+        if not isinstance(cap, int) or isinstance(cap, bool) or cap <= 0:
+            raise LifecycleError("invalid-resource-cap", f"{field} must be a positive integer", {"field": field})
+    require_no_import_secrets(source_path.name)
+    files, root, source_kind = _collection_files(source_path, max_files=max_files)
+    blockers: list[dict[str, Any]] = []
+    if not files:
+        blockers.append({"code": "markdown-collection-empty", "sourceLabel": source_path.name})
+    if len(files) > max_files:
+        blockers.append({"code": "markdown-collection-file-cap-exceeded", "fileCount": len(files), "cap": max_files})
+        files = []
+
+    entries: list[dict[str, Any]] = []
+    combined_parts: list[str] = []
+    total_input_bytes = 0
+    for path in sorted(files):
+        rel = path.relative_to(root).as_posix() if source_kind == "directory" else path.name
+        require_no_import_secrets(rel)
+        with open_authority_read(path) as handle:
+            data = handle.read(max_input_bytes - total_input_bytes + 1)
+        total_input_bytes += len(data)
+        if total_input_bytes > max_input_bytes:
+            blockers.append(
+                {
+                    "code": "markdown-collection-input-cap-exceeded",
+                    "inputBytes": total_input_bytes,
+                    "cap": max_input_bytes,
+                }
+            )
+            break
+        entries.append({"label": rel, "digest": sha256_hex(data), "inputBytes": len(data)})
+        text = _decode_markdown(data, rel, blockers)
+        require_no_import_secrets(text)
+        combined_parts.append(f"# Source: {rel}\n\n{text.strip()}\n")
+    combined_text = "" if blockers else "\n".join(combined_parts).strip()
+
+    body = {
+        "schemaVersion": MARKDOWN_COLLECTION_SCHEMA,
+        "sourceKind": source_kind,
+        "sourceLabel": source_path.name,
+        "ordering": "lexical-relative-posix",
+        "fileCount": len(entries),
+        "totalInputBytes": total_input_bytes,
+        "resourceCaps": {"maxInputBytes": max_input_bytes, "maxFiles": max_files},
+        "files": entries,
+        "blockers": blockers,
+    }
+    body["collectionDigest"] = canonical_digest(body)
+    body["_combinedText"] = combined_text
+    return body
+
+
+def _collection_files(source_path: Path, *, max_files: int) -> tuple[list[Path], Path, str]:
+    discovery = _MarkdownDiscovery(max_files)
+    try:
+        try:
+            info = discovery.operation(source_path.lstat)
+        except FileNotFoundError:
+            return [], source_path.parent, "missing"
+        _require_unlinked(info)
+        if stat.S_ISREG(info.st_mode):
+            return [source_path], source_path.parent, "file"
+        if stat.S_ISDIR(info.st_mode):
+            return discovery.collect(source_path, info), source_path, "directory"
+        return [], source_path.parent, "missing"
+    except OSError:
+        raise LifecycleError(
+            "markdown-collection-discovery-unavailable", "Markdown discovery metadata or directory is unavailable"
+        ) from None
+
+
+def _require_unlinked(info: os.stat_result) -> None:
+    if stat.S_ISLNK(info.st_mode) or getattr(info, "st_file_attributes", 0) & 0x400:
+        raise LifecycleError(
+            "authority-input-not-regular", "Markdown discovery must not traverse links or reparse points"
+        )
+
+
+class _MarkdownDiscovery:
+    """Count all directory entries, with cooperative checks around filesystem I/O."""
+
+    def __init__(self, max_files: int) -> None:
+        self.max_files = max_files
+        self.deadline = time.monotonic() + _DISCOVERY_SECONDS
+        self.entries = 0
+        self.directories = 1  # The source directory is depth zero and counts toward the cap.
+        self.files: list[Path] = []
+        self.pending: list[tuple[Path, int, os.stat_result]] = []
+
+    def check(self) -> None:
+        if time.monotonic() >= self.deadline:
+            self.exceeded("seconds", _DISCOVERY_SECONDS)
+
+    def exceeded(self, limit: str, cap: int | float) -> None:
+        raise LifecycleError(
+            "markdown-collection-discovery-limit-exceeded",
+            "Markdown discovery exceeded its cooperative resource limit",
+            {"limit": limit, "cap": cap},
+        )
+
+    def operation(self, action: Callable[[], _T]) -> _T:
+        self.check()
+        try:
+            return action()
+        finally:
+            self.check()
+
+    def collect(self, root: Path, info: os.stat_result) -> list[Path]:
+        self.pending.append((root, 0, info))
+        while self.pending and len(self.files) <= self.max_files:
+            self.scan(*self.pending.pop())
+        return self.files
+
+    def scan(self, path: Path, depth: int, expected: os.stat_result) -> None:
+        self.check()
+        # Reopen one guarded directory at a time: the queue retains identities,
+        # never thousands of ancestor descriptors. The guard validates bindings
+        # on exit; fd-based scandir stays attached to the held POSIX directory.
+        with hold_authority_directory(path) as directory:
+            self.check()
+            if directory.fd is None:
+                actual = self.operation(directory.path.lstat)
+            else:
+                descriptor = directory.fd
+                actual = self.operation(lambda: os.fstat(descriptor))
+            expected_identity = (expected.st_dev, expected.st_ino, expected.st_mode)
+            actual_identity = (actual.st_dev, actual.st_ino, actual.st_mode)
+            if expected_identity != actual_identity:
+                raise LifecycleError("authority-input-changed", "authority input changed during I/O")
+            self.check()
+            with os.scandir(directory.fd if directory.fd is not None else directory.path) as entries:
+                self.check()
+                while len(self.files) <= self.max_files:
+                    try:
+                        entry = self.operation(lambda: next(entries))
+                    except StopIteration:
+                        break
+                    if self.entries >= _MAX_DISCOVERY_ENTRIES:
+                        # A sentinel detects overflow; do not inspect its metadata.
+                        self.exceeded("entries", _MAX_DISCOVERY_ENTRIES)
+                    self.entries += 1
+                    # DirEntry.stat caches metadata and has zero identity fields
+                    # on Windows 3.11. Use the held directory's fresh no-follow stat.
+                    info = self.operation(partial(directory.stat_child, entry.name))
+                    self.visit(path / entry.name, depth, info)
+                self.check()
+            self.check()
+        self.check()
+
+    def visit(self, path: Path, depth: int, info: os.stat_result) -> None:
+        _require_unlinked(info)
+        if stat.S_ISDIR(info.st_mode):
+            if depth >= _MAX_DISCOVERY_DEPTH:
+                self.exceeded("depth", _MAX_DISCOVERY_DEPTH)
+            if self.directories >= _MAX_DISCOVERY_DIRECTORIES:
+                self.exceeded("directories", _MAX_DISCOVERY_DIRECTORIES)
+            self.directories += 1
+            self.pending.append((path, depth + 1, info))
+        elif fnmatch(path.name, "*.md"):
+            if not stat.S_ISREG(info.st_mode):
+                raise LifecycleError("authority-input-not-regular", "Markdown input must be a regular file")
+            self.files.append(path)
+
+
+def _decode_markdown(data: bytes, label: str, blockers: list[dict[str, Any]]) -> str:
+    try:
+        return data.decode("utf-8")
+    except UnicodeDecodeError:
+        blockers.append({"code": "markdown-collection-decode-failed", "sourceLabel": label})
+        return ""

{agent_lifecycle_kit-2.25.0 → agent_lifecycle_kit-2.26.0/src/agent_lifecycle_kit.egg-info}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: agent-lifecycle-kit
-Version: 2.25.0
+Version: 2.26.0
 Summary: Provider-neutral lifecycle control layer that helps coding agents plan, execute, validate, and prove software tasks to completion.
 Author: Agent Lifecycle Kit contributors
 License: Apache-2.0
@@ -41,7 +41,7 @@ ALK provides a verifiable finish for external-agent work with proportionate cont
 ## Quick start
 Start with [install ALK and make the first run](docs/guides/install-and-first-run.md). Release 2.23 adds [compact S0/S1 authoring](docs/reference/tier-proportional-plan-packages.md) with explicit checklist completion and unchanged approval gates. Release 2.22 adds [conservative plan-review reuse](docs/reference/conservative-delta-review.md) with independent current applicability and no automatic full reruns. Release 2.21 adds [command accounting, bounded preparation and response recovery](docs/reference/source-native-workflow-accounting.md). Release 2.20 adds [target-bound initialization](docs/reference/cli.md#release-220-composed-initialization-and-digest-display), [bounded local continuation](docs/reference/workflow-continuation.md) and batched Git snapshots. Release 2.19 adds [safe artifact authoring and planning preflight](docs/reference/safe-artifact-authoring.md), [bounded review handoff](docs/reference/review-mesh.md#bounded-review-handoff) and thin skills with complete references. Release 2.18 bounds process/Git diagnostics, hardens signing and distinguishes OpenCode 1.x/2.x compatibility ([adapter notes](docs/adapters/opencode.md)); 2.17 strengthens [workflow validation and recovery](docs/reference/workflow-completeness.md). Release 2.16 introduced cumulative accounting within verified host capabilities; see [retention maintenance](docs/operations/retention-maintenance.md).
 
-It includes macOS, Linux and Windows instructions, Python 3.11-3.14, [PyPI](https://pypi.org/project/agent-lifecycle-kit/) installation with `python -m pip install agent-lifecycle-kit==2.25.0`, and common `agent-lifecycle version` errors. See [2.15 migration notes](docs/reference/cli.md#migrating-authority-artifacts-to-215) and [authority boundaries](docs/reference/authority-artifact-boundaries.md).
+It includes macOS, Linux and Windows instructions, Python 3.11-3.14, [PyPI](https://pypi.org/project/agent-lifecycle-kit/) installation with `python -m pip install agent-lifecycle-kit==2.26.0`, and common `agent-lifecycle version` errors. See [2.15 migration notes](docs/reference/cli.md#migrating-authority-artifacts-to-215) and [authority boundaries](docs/reference/authority-artifact-boundaries.md).
 
 For a quick start, run:
 
@@ -194,9 +194,9 @@ Adapter installation and support-level details live in [Adapter install](docs/ad
 - Release assets: [Task templates](templates/tasks/README.md), [Neutrality scanning](docs/reference/neutrality.md), and [Release security](docs/security/release-security.md).
 Apache-2.0. See [LICENSE](LICENSE).
 
-## Release 2.25: focused security and maintainability
+## Release 2.26: bounded discovery and precise quality gates
 
-Release 2.25 hardens sensitive imports and launcher checks, repairs named ownership seams, and checks untyped Python function bodies with mypy. See the [current security
-controls](docs/security/release-security.md) and [architecture map](docs/architecture/system-architecture.md). The [staged reading
-pilot](docs/guides/controlled-technical-english.md) preserves full references and approval gates; byte/word measurements do not prove token savings. Existing OpenCode 1.x/2.x
-qualification and unsupported host routes remain unchanged. Local preparation, exact-candidate CI, native final proof and public publication are separate gates.
+Markdown imports bound examined entries, directories, depth and cooperative elapsed time; see the [security control map](docs/security/release-security.md). Tracked-source checks
+reuse the existing bare-token families. [Quality controls](docs/architecture/system-architecture.md)
+require both combined and statement coverage of 76 percent and retain legacy policy interpretation. The existing [reading pilot](docs/guides/controlled-technical-english.md) and host limits
+remain; measurements do not establish token savings or host enforcement.

agent_lifecycle_kit-2.25.0/src/agent_lifecycle/imports/multi_markdown.py

@@ -1,127 +0,0 @@
-"""Deterministic Markdown file and folder import helpers."""
-
-from __future__ import annotations
-
-from itertools import islice
-from pathlib import Path
-from typing import Any
-
-from agent_lifecycle.contracts import LifecycleError, canonical_digest, sha256_hex
-from agent_lifecycle.contracts.authority_io import open_authority_read
-from agent_lifecycle.imports.planning import (
-    DEFAULT_MAX_INPUT_BYTES,
-    DEFAULT_TARGET_TOKENS,
-    import_planning_text,
-    require_no_import_secrets,
-)
-
-MARKDOWN_COLLECTION_SCHEMA = "agent-markdown-source-collection.v1"
-DEFAULT_MAX_MARKDOWN_FILES = 32
-
-
-def import_markdown_collection(
-    source_path: Path,
-    *,
-    package_id: str = "markdown-import",
-    max_input_bytes: int = DEFAULT_MAX_INPUT_BYTES,
-    target_tokens: int = DEFAULT_TARGET_TOKENS,
-    max_files: int = DEFAULT_MAX_MARKDOWN_FILES,
-    dialect_profile: dict[str, Any] | None = None,
-) -> dict[str, Any]:
-    """Import one Markdown file or a directory of Markdown files as draft input."""
-
-    collection = collect_markdown_collection(source_path, max_input_bytes=max_input_bytes, max_files=max_files)
-    text = collection.pop("_combinedText", "")
-    result = import_planning_text(
-        text,
-        source_label=collection["sourceLabel"],
-        package_id=package_id,
-        max_input_bytes=max_input_bytes,
-        target_tokens=target_tokens,
-        dialect_profile=dialect_profile,
-    )
-    result["markdownCollection"] = collection
-    if collection["blockers"]:
-        result["blockers"].extend(collection["blockers"])
-        result["status"] = "FAIL"
-        result["candidateLifecycleStatus"] = "BLOCKED"
-        result["candidatePlan"] = None
-    result["importDigest"] = canonical_digest({key: value for key, value in result.items() if key != "importDigest"})
-    return result
-
-
-def collect_markdown_collection(
-    source_path: Path,
-    *,
-    max_input_bytes: int = DEFAULT_MAX_INPUT_BYTES,
-    max_files: int = DEFAULT_MAX_MARKDOWN_FILES,
-) -> dict[str, Any]:
-    for field, cap in (("maxFiles", max_files), ("maxInputBytes", max_input_bytes)):
-        if not isinstance(cap, int) or isinstance(cap, bool) or cap <= 0:
-            raise LifecycleError("invalid-resource-cap", f"{field} must be a positive integer", {"field": field})
-    require_no_import_secrets(source_path.name)
-    files, root, source_kind = _collection_files(source_path, max_files=max_files)
-    blockers: list[dict[str, Any]] = []
-    if not files:
-        blockers.append({"code": "markdown-collection-empty", "sourceLabel": source_path.name})
-    if len(files) > max_files:
-        blockers.append({"code": "markdown-collection-file-cap-exceeded", "fileCount": len(files), "cap": max_files})
-        files = []
-
-    entries: list[dict[str, Any]] = []
-    combined_parts: list[str] = []
-    total_input_bytes = 0
-    for path in sorted(files):
-        rel = path.relative_to(root).as_posix() if source_kind == "directory" else path.name
-        require_no_import_secrets(rel)
-        with open_authority_read(path) as handle:
-            data = handle.read(max_input_bytes - total_input_bytes + 1)
-        total_input_bytes += len(data)
-        if total_input_bytes > max_input_bytes:
-            blockers.append(
-                {
-                    "code": "markdown-collection-input-cap-exceeded",
-                    "inputBytes": total_input_bytes,
-                    "cap": max_input_bytes,
-                }
-            )
-            break
-        entries.append({"label": rel, "digest": sha256_hex(data), "inputBytes": len(data)})
-        text = _decode_markdown(data, rel, blockers)
-        require_no_import_secrets(text)
-        combined_parts.append(f"# Source: {rel}\n\n{text.strip()}\n")
-    combined_text = "" if blockers else "\n".join(combined_parts).strip()
-
-    body = {
-        "schemaVersion": MARKDOWN_COLLECTION_SCHEMA,
-        "sourceKind": source_kind,
-        "sourceLabel": source_path.name,
-        "ordering": "lexical-relative-posix",
-        "fileCount": len(entries),
-        "totalInputBytes": total_input_bytes,
-        "resourceCaps": {"maxInputBytes": max_input_bytes, "maxFiles": max_files},
-        "files": entries,
-        "blockers": blockers,
-    }
-    body["collectionDigest"] = canonical_digest(body)
-    body["_combinedText"] = combined_text
-    return body
-
-
-def _collection_files(source_path: Path, *, max_files: int) -> tuple[list[Path], Path, str]:
-    if source_path.is_symlink():
-        raise LifecycleError("authority-input-not-regular", "import source must not be a symlink")
-    if source_path.is_file():
-        return [source_path], source_path.parent, "file"
-    if source_path.is_dir():
-        candidates = (path for path in source_path.rglob("*.md") if path.is_file() or path.is_symlink())
-        return list(islice(candidates, max_files + 1)), source_path, "directory"
-    return [], source_path.parent, "missing"
-
-
-def _decode_markdown(data: bytes, label: str, blockers: list[dict[str, Any]]) -> str:
-    try:
-        return data.decode("utf-8")
-    except UnicodeDecodeError:
-        blockers.append({"code": "markdown-collection-decode-failed", "sourceLabel": label})
-        return ""