Implements parallel ACMS indexing infrastructure to meet the 60-second
performance target for 10,000-file projects on modern hardware.

Key components added:
- ACMSIndexingConfig: configurable workers (min(32, cpu_count+4)), batch_size,
  max_file_size_bytes, and progress_interval settings
- FileFilter: early-stage binary detection (null-byte heuristic), size threshold
  enforcement, and .acmsignore/.gitignore pattern exclusion
- IndexCache: on-disk JSON cache keyed by path+mtime+size with atomic file
  replacement to avoid corruption on crash
- IndexProgressReporter: thread-safe progress tracking with configurable
  time/file-count intervals and callback support
- ParallelIndexer: ThreadPoolExecutor-based concurrent file hashing with
  streaming batch support; per-file exceptions caught without aborting the run
- parallel_walk_and_index: drop-in replacement for walk_and_index integrating
  all components

Also adds I001 to per-file-ignores for features/steps/*.py to allow the
established pattern of lazy imports inside step functions.

Closes #9330

@@ -0,0 +7,4 @@
+* Tokens mode truncation
+* System prompt preservation
+* Budget enforcement
+* can_handle confidence

@@ -0,0 +2,4 @@
+ACMS (Agent Configuration Management System) - Parallel File Indexing Module
+
+Optimized for projects with 10,000+ files using parallel processing,
+incremental indexing, and efficient caching.

@@ -0,0 +438,4 @@
+        base_dir: Path | str,
+        worker_count: int = DEFAULT_WORKER_COUNT,
+        hash_algorithm: str = "sha256",
+        cache_path: str = DEFAULT_CACHE_PATH,

@@ -0,0 +551,4 @@
+        Args:
+            force: If True, ignore cached index and rebuild completely.
+            on_progress: Optional callback receiving IndexProgress updates.
+            max_workers: Override worker count for this operation.

@@ -0,0 +645,4 @@
+                "failed": failed_count,
+                "errors": errors[:50],
+            })
+

@@ -0,0 +867,4 @@
+        """Export the full index as a dictionary."""
+        output_path = output_path or self.cache_path
+        export_data = {
+            "base_dir": str(self.base_dir),

@@ -0,0 +91,4 @@
+
+    # -- timing methods -------------------------------------------------------
+
+    def time_parallel_index(self, file_count: int) -> None:

@@ -0,0 +80,4 @@
+def _index_single_file(
+    file_path_str: str,
+    base_dir_str: str,
+    hash_algorithm: str = "md5",

@@ -0,0 +89,4 @@
+        return dict(self._index)
+
+    def _load_cache(self) -> bool:
+        """Load index from cache file. Returns True on success."""

@@ -0,0 +307,4 @@
+                    str(fp),
+                    str(self.base_dir),
+                    self.hash_algorithm,
+                ): fp

@@ -0,0 +215,4 @@
+                    break
+                fp = future_to_file[future]
+                try:
+                    result = future.result(timeout=30)

@@ -0,0 +20,4 @@
+The strategy is registered in the ``ACMSPipeline`` under the key
+``"sliding_window"`` so it can be selected by name.
+
+Based on ``docs/specification.md`` §25207 and issue #9995.

@@ -342,0 +687,4 @@
+
+        return ordered_entries
+
+    def _save_cache(self) -> None:

@@ -235,6 +235,15 @@ Changed `wf10_batch.robot` to be less likely to create files, and
 ### Changed
 
 - Fixed stale `AUTO-BUG-POOL` tracking prefix references in automation-tracking.md documentation and agent-system-specification.md spec document, replaced with correct `AUTO-BUG-SUP` prefix used by the bug-hunt-pool-supervisor agent (#7875).
+=======

@@ -0,0 +7,4 @@
+  reduces indexing wall-clock time compared to sequential chunking."
+
+  Background:
+    The parallel engine uses configurable worker threads (max_workers)

@@ -0,0 +145,4 @@
+    context.add_cleanup(shutil.rmtree, d, True)
+
+
+@given(

@@ -67,7 +77,7 @@ class IndexEntry(BaseModel):
     modified_at: datetime

@@ -107,0 +175,4 @@
+                "files_processed": self.files_processed,
+                "files_skipped": self.files_skipped,
+                "files_indexed": self.files_indexed,
+                "bytes_read": self.bytes_read,

@@ -342,0 +629,4 @@
+            self.progress.record_processed()
+            return None
+
+        try:

@@ -0,0 +1,71 @@
+@acms-parallel-indexing

@@ -857,6 +857,17 @@ def benchmark_regression(session: nox.Session):
     """Run Airspeed Velocity benchmarks regression test."""
     session.install("-e", ".[tests]")
     config_path = "asv.conf.json"
+

@@ -342,0 +594,4 @@
+        Performs stat + optional binary detection.  Returns an IndexEntry
+        on success or None if the file should be skipped.
+
+        This function is designed to run inside ThreadPoolExecutor without