test: refresh oversized split plan

mcp_servers/memory_server.py

@@ -6,7 +6,6 @@ Imports MemoryManager and MemoryVectorStore from the Odysseus codebase.
 """
 
 import asyncio
-import os
 import sys
 import time
 from pathlib import Path
@@ -24,55 +23,6 @@ _memory_manager = None
 _memory_vector = None
 _initialized = False
 
-_OWNER_ENV_KEYS = ("ODYSSEUS_MCP_MEMORY_OWNER", "ODYSSEUS_MEMORY_OWNER")
-_OWNER_SCOPE_ERROR = (
-    "Error: Memory MCP owner is not configured for an owner-scoped memory store. "
-    "Set ODYSSEUS_MCP_MEMORY_OWNER for this server or use the owner-aware native memory tool."
-)
-
-
-def _configured_owner() -> str | None:
-    for key in _OWNER_ENV_KEYS:
-        owner = os.environ.get(key, "").strip()
-        if owner:
-            return owner
-    return None
-
-
-def _entry_owner(entry: dict) -> str | None:
-    owner = entry.get("owner")
-    if owner is None:
-        return None
-    owner_text = str(owner).strip()
-    return owner_text or None
-
-
-def _owner_scoped_store(entries: list[dict]) -> bool:
-    return any(_entry_owner(entry) for entry in entries if isinstance(entry, dict))
-
-
-def _scope_entries() -> tuple[str | None, list[dict], list[dict], str | None]:
-    """Return configured owner, all entries, visible entries, and optional error."""
-    entries = _memory_manager.load_all()
-    owner = _configured_owner()
-    if owner is None and _owner_scoped_store(entries):
-        return None, entries, [], _OWNER_SCOPE_ERROR
-    if owner is None:
-        visible = [
-            entry for entry in entries
-            if isinstance(entry, dict) and _entry_owner(entry) is None
-        ]
-    else:
-        visible = [
-            entry for entry in entries
-            if isinstance(entry, dict) and _entry_owner(entry) == owner
-        ]
-    return owner, entries, visible, None
-
-
-def _text_result(text: str) -> list[TextContent]:
-    return [TextContent(type="text", text=text)]
-
 
 def _ensure_init():
     """Lazy-init memory managers on first use."""
@@ -125,26 +75,24 @@ async def list_tools() -> list[Tool]:
 @server.call_tool()
 async def call_tool(name: str, arguments: dict) -> list[TextContent]:
     if name != "manage_memory":
-        return _text_result(f"Unknown tool: {name}")
+        return [TextContent(type="text", text=f"Unknown tool: {name}")]
 
     _ensure_init()
     if not _memory_manager:
-        return _text_result("Error: Memory manager not available")
+        return [TextContent(type="text", text="Error: Memory manager not available")]
 
     action = arguments.get("action", "")
 
     if action == "list":
         category_filter = arguments.get("category", "")
-        _owner, _all_memories, memories, scope_error = _scope_entries()
-        if scope_error:
-            return _text_result(scope_error)
+        memories = _memory_manager.load()
         if category_filter:
             memories = [m for m in memories if m.get("category", "").lower() == category_filter.lower()]
         if not memories:
             msg = "No memories found"
             if category_filter:
                 msg += f" in category '{category_filter}'"
-            return _text_result(msg + ".")
+            return [TextContent(type="text", text=msg + ".")]
 
         lines = [f"Found {len(memories)} memory entries:\n"]
         for m in memories:
@@ -154,17 +102,15 @@ async def call_tool(name: str, arguments: dict) -> list[TextContent]:
             if len(text) > 150:
                 text = text[:150] + "..."
             lines.append(f"- [{cat}] `{mid}` — {text}")
-        return _text_result("\n".join(lines))
+        return [TextContent(type="text", text="\n".join(lines))]
 
     elif action == "add":
         text = arguments.get("text", "")
         category = arguments.get("category", "fact")
         if not text:
-            return _text_result("Error: Memory text cannot be empty")
-        owner, memories, _visible, scope_error = _scope_entries()
-        if scope_error:
-            return _text_result(scope_error)
-        entry = _memory_manager.add_entry(text, source="ai_agent", category=category, owner=owner)
+            return [TextContent(type="text", text="Error: Memory text cannot be empty")]
+        entry = _memory_manager.add_entry(text, source="ai_agent", category=category)
+        memories = _memory_manager.load_all()
         memories.append(entry)
         _memory_manager.save(memories)
         if _memory_vector and _memory_vector.healthy:
@@ -172,28 +118,25 @@ async def call_tool(name: str, arguments: dict) -> list[TextContent]:
                 _memory_vector.add(entry["id"], text)
             except Exception:
                 pass
-        return _text_result(f"Memory added: [{category}] {text} (id: {entry['id'][:8]})")
+        return [TextContent(type="text", text=f"Memory added: [{category}] {text} (id: {entry['id'][:8]})")]
 
     elif action == "edit":
         memory_id = arguments.get("memory_id", "")
         new_text = arguments.get("text", "")
         if not memory_id or not new_text:
-            return _text_result("Error: edit needs memory_id and text")
-        _owner, memories, visible, scope_error = _scope_entries()
-        if scope_error:
-            return _text_result(scope_error)
+            return [TextContent(type="text", text="Error: edit needs memory_id and text")]
+        memories = _memory_manager.load_all()
+        found = False
         full_id = None
-        for m in visible:
-            if m.get("id", "").startswith(memory_id):
-                full_id = m["id"]
-                break
-        if not full_id:
-            return _text_result(f"Error: Memory '{memory_id}' not found")
         for m in memories:
-            if m.get("id") == full_id:
+            if m.get("id", "").startswith(memory_id):
                 m["text"] = new_text
                 m["timestamp"] = int(time.time())
+                found = True
+                full_id = m["id"]
                 break
+        if not found:
+            return [TextContent(type="text", text=f"Error: Memory '{memory_id}' not found")]
         _memory_manager.save(memories)
         if _memory_vector and _memory_vector.healthy and full_id:
             try:
@@ -201,26 +144,24 @@ async def call_tool(name: str, arguments: dict) -> list[TextContent]:
                 _memory_vector.add(full_id, new_text)
             except Exception:
                 pass
-        return _text_result(f"Memory updated: {new_text}")
+        return [TextContent(type="text", text=f"Memory updated: {new_text}")]
 
     elif action == "delete":
         memory_id = arguments.get("memory_id", "")
         if not memory_id:
-            return _text_result("Error: delete needs memory_id")
-        _owner, memories, visible, scope_error = _scope_entries()
-        if scope_error:
-            return _text_result(scope_error)
+            return [TextContent(type="text", text="Error: delete needs memory_id")]
+        memories = _memory_manager.load_all()
         full_id = None
         deleted_text = ""
         deleted_category = ""
-        for m in visible:
+        for m in memories:
             if m.get("id", "").startswith(memory_id):
                 full_id = m["id"]
                 deleted_text = m.get("text", "")
                 deleted_category = m.get("category", "")
                 break
         if not full_id:
-            return _text_result(f"Error: Memory '{memory_id}' not found")
+            return [TextContent(type="text", text=f"Error: Memory '{memory_id}' not found")]
         memories = [m for m in memories if m.get("id") != full_id]
         _memory_manager.save(memories)
         if _memory_vector and _memory_vector.healthy and full_id:
@@ -230,32 +171,30 @@ async def call_tool(name: str, arguments: dict) -> list[TextContent]:
                 pass
         cat = f"[{deleted_category}] " if deleted_category else ""
         snippet = deleted_text if len(deleted_text) <= 120 else deleted_text[:117] + "..."
-        return _text_result(f"Memory deleted: {cat}{snippet} (id: {memory_id})")
+        return [TextContent(type="text", text=f"Memory deleted: {cat}{snippet} (id: {memory_id})")]
 
     elif action == "search":
         query = arguments.get("text", "")
         if not query:
-            return _text_result("Error: search needs text (query)")
-        _owner, _all_memories, memories, scope_error = _scope_entries()
-        if scope_error:
-            return _text_result(scope_error)
+            return [TextContent(type="text", text="Error: search needs text (query)")]
+        memories = _memory_manager.load()
         if hasattr(_memory_manager, 'get_relevant_memories'):
             results = _memory_manager.get_relevant_memories(query, memories, threshold=0.05, max_items=20)
         else:
             query_lower = query.lower()
             results = [m for m in memories if query_lower in m.get("text", "").lower()][:20]
         if not results:
-            return _text_result(f"No memories found matching '{query}'.")
+            return [TextContent(type="text", text=f"No memories found matching '{query}'.")]
         lines = [f"Found {len(results)} matching memories:\n"]
         for m in results:
             cat = m.get("category", "fact")
             mid = m.get("id", "?")[:8]
             text = m.get("text", "")
             lines.append(f"- [{cat}] `{mid}` — {text}")
-        return _text_result("\n".join(lines))
+        return [TextContent(type="text", text="\n".join(lines))]
 
     else:
-        return _text_result(f"Error: Unknown action '{action}'. Use: list, add, edit, delete, search")
+        return [TextContent(type="text", text=f"Error: Unknown action '{action}'. Use: list, add, edit, delete, search")]
 
 
 async def run():

services/search/content.py

@@ -15,7 +15,7 @@ from urllib.parse import urljoin, urlparse
 import httpx
 from bs4 import BeautifulSoup
 
-from src.constants import WEB_FETCH_SOFT_MAX_BYTES, WEB_FETCH_HARD_MAX_BYTES, WEB_FETCH_USER_AGENT
+from src.constants import WEB_FETCH_SOFT_MAX_BYTES, WEB_FETCH_HARD_MAX_BYTES
 
 from .analytics import RateLimitError, error_logger
 from .cache import (
@@ -369,7 +369,7 @@ def fetch_webpage_content(url: str, timeout: int = 5, retry_attempt: int = 0,
     # Fetch
     try:
         headers = {
-            "User-Agent": WEB_FETCH_USER_AGENT,
+            "User-Agent": "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/91.0.4472.124 Safari/537.36",
             "Accept": "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8",
             "Accept-Language": "en-US,en;q=0.5",
             # identity so the streamed size cap in _get_public_url stays honest

services/search/providers.py

@@ -9,7 +9,7 @@ from urllib.parse import urljoin, urlparse, parse_qs
 import httpx
 from bs4 import BeautifulSoup
 
-from src.constants import SEARXNG_INSTANCE, REQUEST_TIMEOUT, WEB_FETCH_USER_AGENT
+from src.constants import SEARXNG_INSTANCE, REQUEST_TIMEOUT
 from .analytics import RateLimitError, error_logger
 from .query import build_enhanced_query
 
@@ -138,7 +138,7 @@ def searxng_search_api(query: str, count: Optional[int] = None, categories: str
     count = count if count is not None else _get_result_count()
     instance = _get_search_instance()
     api_key = ""
-    headers = {"User-Agent": WEB_FETCH_USER_AGENT}
+    headers = {"User-Agent": "Mozilla/5.0"}
     if api_key:
         headers["Authorization"] = f"Bearer {api_key}"
     # News/fresh queries do badly in the 'general' category — it favours
@@ -250,7 +250,7 @@ def searxng_search(query, max_results=10):
     """Search using SearXNG instance - parsing HTML."""
     instance = _get_search_instance()
     api_key = ""
-    req_headers = {"User-Agent": WEB_FETCH_USER_AGENT}
+    req_headers = {"User-Agent": "Mozilla/5.0"}
     if api_key:
         req_headers["Authorization"] = f"Bearer {api_key}"
     try:
@@ -389,7 +389,7 @@ def duckduckgo_search(query: str, count: Optional[int] = None, time_filter: Opti
             response = httpx.get(
                 "https://html.duckduckgo.com/html/",
                 params={"q": query, "kp": _safesearch_for("duckduckgo_html")},
-                headers={"User-Agent": WEB_FETCH_USER_AGENT},
+                headers={"User-Agent": "Mozilla/5.0"},
                 timeout=REQUEST_TIMEOUT,
             )
             response.raise_for_status()

src/constants.py

@@ -57,13 +57,7 @@ MEMORY_VECTORS_DIR = os.path.join(DATA_DIR, "memory_vectors")
 
 # Paths with an intentional dedicated env override, defaulting under DATA_DIR.
 MAIL_ATTACHMENTS_DIR = os.getenv("ODYSSEUS_MAIL_ATTACHMENTS_DIR", os.path.join(DATA_DIR, "mail-attachments"))
-# `or` (not os.getenv's default arg) so a PRESENT-but-EMPTY value falls back to
-# the default. docker-compose.yml injects `FASTEMBED_CACHE_PATH=${FASTEMBED_CACHE_PATH:-}`,
-# which sets the var to "" when the host hasn't defined it. os.getenv(name, default)
-# only returns the default when the var is ABSENT, so the empty string would win →
-# os.makedirs("") raises [Errno 2] No such file or directory: '' → FastEmbed fails to
-# init and all vector features (RAG, semantic memory, tool index) silently degrade.
-FASTEMBED_CACHE_DIR = os.getenv("FASTEMBED_CACHE_PATH") or os.path.join(DATA_DIR, "fastembed_cache")
+FASTEMBED_CACHE_DIR = os.getenv("FASTEMBED_CACHE_PATH", os.path.join(DATA_DIR, "fastembed_cache"))
 
 # Agent tool output limits (single source of truth — imported by tool_execution.py,
 # tool_implementations.py, agent_tools.py, and any other module that needs them)
@@ -84,13 +78,6 @@ MAX_CONTEXT_MESSAGES = 90
 REQUEST_TIMEOUT = 20
 OPENAI_COMPAT_PATH = "/v1/chat/completions"
 
-# Outbound UA for web_fetch / web_search scraping; common desktop UA so pages serve normal HTML.
-WEB_FETCH_USER_AGENT = os.environ.get(
-    "WEB_FETCH_USER_AGENT",
-    "Mozilla/5.0 (Windows NT 10.0; Win64; x64) AppleWebKit/537.36 "
-    "(KHTML, like Gecko) Chrome/148.0.0.0 Safari/537.36",
-)
-
 # Environment variables with defaults
 DEFAULT_HOST = os.getenv("LLM_HOST", "localhost")
 LLM_HOSTS = [h.strip() for h in os.getenv("LLM_HOSTS", "").split(",") if h.strip()]

src/task_scheduler.py

@@ -236,29 +236,6 @@ def _digest_windows(now):
     ]
 
 
-def _checkin_calendar_events(db, owner, start, end):
-    """Calendar events in [start, end] for ONE owner, for the check-in digest.
-
-    Ownership lives on CalendarCal.owner; events inherit it via calendar_id.
-    The digest query had no owner scope, so it pulled EVERY user's events into
-    one user's check-in (a cross-tenant leak of summaries/locations). Scope it
-    by joining CalendarCal, mirroring routes/calendar_routes.list_events.
-    """
-    from core.database import CalendarEvent as _CE, CalendarCal as _CC
-    return (
-        db.query(_CE)
-        .join(_CC, _CE.calendar_id == _CC.id)
-        .filter(
-            _CC.owner == owner,
-            _CE.dtstart >= start,
-            _CE.dtstart <= end,
-            _CE.status != "cancelled",
-        )
-        .order_by(_CE.dtstart)
-        .all()
-    )
-
-
 class TaskScheduler:
     def __init__(self, session_manager):
         self._session_manager = session_manager
@@ -1150,7 +1127,11 @@ class TaskScheduler:
                     # Strip timezone for naive DB comparison
                     _s = start.replace(tzinfo=None) if start.tzinfo else start
                     _e = end.replace(tzinfo=None) if end.tzinfo else end
-                    evs = _checkin_calendar_events(_db, task.owner, _s, _e)
+                    evs = _db.query(_CE).filter(
+                        _CE.dtstart >= _s,
+                        _CE.dtstart <= _e,
+                        _CE.status != "cancelled",
+                    ).order_by(_CE.dtstart).all()
                     if not evs:
                         continue
                     # Group by importance for richer output

tests/OVERSIZED_TEST_SPLIT_PLAN.md

@@ -0,0 +1,326 @@
+# Oversized Test File Split Plan
+
+## Purpose
+
+This document plans future oversized test-file splits using current repo data.
+It does not move files, rewrite assertions, extract helpers, or change CI.
+
+## Roadmap context
+
+- Issue: #3983
+- Parent tracker: #2523
+- Follows #3973 / #3982, the report-only order-sensitivity diagnostics slice.
+
+## Methodology
+
+Metrics were generated from the current test tree using:
+
+- physical line counts for every recursive `test_*.py` file under `tests/`;
+- AST counts for `test_*` functions and `Test*` classes;
+- one `pytest --collect-only -q tests` run to count collected items per file;
+- current taxonomy classification from `tests._taxonomy.classify_test_path`; and
+- static setup-signal scans for route/API, DB/session, import-state, security, filesystem, subprocess/script, async/threading, and UI/static indicators.
+
+Static signals are not proof of risk. They are review prompts.
+Future split PRs must still inspect each file manually before editing.
+
+## Current summary
+
+- test files scanned: 583
+- collected pytest items counted: 3586
+- large-file threshold: 300 lines
+- large-collected threshold: 20 collected items
+
+Area distribution:
+
+| Value | Files |
+|---|---:|
+| cli | 28 |
+| helpers | 1 |
+| js | 39 |
+| routes | 23 |
+| security | 77 |
+| services | 144 |
+| uncategorized | 234 |
+| unit | 37 |
+
+Sub-area distribution:
+
+| Value | Files |
+|---|---:|
+| api | 6 |
+| atomic | 3 |
+| auth | 9 |
+| calendar | 10 |
+| cli | 28 |
+| confinement | 7 |
+| cookbook | 13 |
+| document | 11 |
+| email | 12 |
+| embedding | 3 |
+| gallery | 5 |
+| history | 3 |
+| js | 39 |
+| llm | 16 |
+| mcp | 8 |
+| memory | 15 |
+| nondict | 7 |
+| nonstring | 22 |
+| owner | 14 |
+| owner_scope | 23 |
+| parse | 4 |
+| provider | 6 |
+| research | 16 |
+| route | 6 |
+| routes | 9 |
+| scheduler | 3 |
+| scope | 5 |
+| security | 9 |
+| session | 16 |
+| ssrf | 3 |
+| webhook | 3 |
+| xss | 5 |
+
+Values below 2 files: 244 values covering 244 files.
+
+## Top files by collected pytest items
+
+| File | Lines | Collected tests | Test defs | Test classes | Area | Sub-area | Signals |
+|---|---:|---:|---:|---:|---|---|---|
+| `tests/test_model_routes.py` | 1778 | 139 | 116 | 10 | routes | routes | route/api, db/session, import-state, async/threading |
+| `tests/test_security_regressions.py` | 1224 | 92 | 68 | 0 | security | security | route/api, db/session, import-state, security, filesystem, async/threading, ui/static |
+| `tests/test_provider_classification.py` | 188 | 67 | 21 | 4 | services | provider | - |
+| `tests/test_cookbook_helpers.py` | 912 | 65 | 65 | 0 | services | cookbook | route/api, filesystem, subprocess/script, async/threading, ui/static |
+| `tests/test_shell_routes.py` | 481 | 63 | 48 | 8 | routes | routes | route/api, import-state, filesystem |
+| `tests/test_pr_blocker_audit.py` | 964 | 58 | 58 | 0 | uncategorized | pr_blocker_audit | import-state, security, filesystem |
+| `tests/test_provider_endpoints.py` | 241 | 58 | 18 | 1 | services | provider | subprocess/script |
+| `tests/test_agent_loop.py` | 469 | 52 | 52 | 5 | uncategorized | agent_loop | db/session, import-state |
+| `tests/test_service_health.py` | 472 | 47 | 42 | 0 | uncategorized | service_health | async/threading |
+| `tests/test_run_focus.py` | 399 | 47 | 44 | 0 | uncategorized | run_focus | security, filesystem, subprocess/script, ui/static |
+| `tests/test_llm_core_temperature.py` | 196 | 41 | 17 | 0 | services | llm | - |
+| `tests/test_endpoint_probing.py` | 411 | 34 | 30 | 6 | uncategorized | endpoint_probing | route/api, db/session, import-state |
+| `tests/test_llm_core_anthropic_temp_omit.py` | 94 | 32 | 6 | 0 | services | llm | db/session |
+| `tests/test_chat_helpers.py` | 264 | 31 | 18 | 0 | uncategorized | chat_helpers | route/api |
+| `tests/test_provider_detection.py` | 148 | 31 | 31 | 5 | services | provider | - |
+| `tests/test_model_context.py` | 251 | 30 | 30 | 4 | uncategorized | model_context | db/session, import-state |
+| `tests/test_endpoint_resolver.py` | 148 | 30 | 30 | 6 | uncategorized | endpoint_resolver | - |
+| `tests/test_embedding_lanes.py` | 1104 | 29 | 29 | 0 | services | embedding | filesystem |
+| `tests/test_upload_limits_centralized.py` | 110 | 29 | 5 | 0 | uncategorized | upload_limits_centralized | import-state, filesystem |
+| `tests/test_email_oauth.py` | 580 | 28 | 25 | 0 | services | email | route/api, db/session, security, async/threading |
+| `tests/test_review_regressions.py` | 930 | 26 | 26 | 0 | uncategorized | review_regressions | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_rename_user_owner_sync.py` | 686 | 26 | 26 | 0 | security | owner | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_helpers_import_state.py` | 426 | 26 | 26 | 0 | helpers | helpers | route/api, db/session, import-state |
+| `tests/test_taxonomy.py` | 145 | 26 | 16 | 0 | uncategorized | taxonomy | security, ui/static |
+| `tests/test_tool_path_confinement.py` | 282 | 24 | 24 | 0 | security | confinement | import-state, filesystem, async/threading |
+| `tests/test_copilot.py` | 170 | 23 | 16 | 0 | uncategorized | copilot | - |
+| `tests/test_research_utils.py` | 97 | 23 | 23 | 2 | services | research | - |
+| `tests/test_api_chat_security.py` | 401 | 22 | 8 | 0 | security | security | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_tool_support_heuristic.py` | 166 | 22 | 22 | 3 | uncategorized | tool_support_heuristic | - |
+| `tests/test_platform_compat.py` | 318 | 21 | 21 | 0 | uncategorized | platform_compat | import-state, filesystem, subprocess/script |
+
+## Top files by physical line count
+
+| File | Lines | Collected tests | Test defs | Test classes | Area | Sub-area | Signals |
+|---|---:|---:|---:|---:|---|---|---|
+| `tests/test_model_routes.py` | 1778 | 139 | 116 | 10 | routes | routes | route/api, db/session, import-state, async/threading |
+| `tests/test_security_regressions.py` | 1224 | 92 | 68 | 0 | security | security | route/api, db/session, import-state, security, filesystem, async/threading, ui/static |
+| `tests/test_embedding_lanes.py` | 1104 | 29 | 29 | 0 | services | embedding | filesystem |
+| `tests/test_pr_blocker_audit.py` | 964 | 58 | 58 | 0 | uncategorized | pr_blocker_audit | import-state, security, filesystem |
+| `tests/test_review_regressions.py` | 930 | 26 | 26 | 0 | uncategorized | review_regressions | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_cookbook_helpers.py` | 912 | 65 | 65 | 0 | services | cookbook | route/api, filesystem, subprocess/script, async/threading, ui/static |
+| `tests/test_rename_user_owner_sync.py` | 686 | 26 | 26 | 0 | security | owner | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_email_oauth.py` | 580 | 28 | 25 | 0 | services | email | route/api, db/session, security, async/threading |
+| `tests/test_api_token_routes.py` | 578 | 17 | 17 | 0 | routes | api_routes | route/api, db/session, import-state, async/threading |
+| `tests/test_shell_routes.py` | 481 | 63 | 48 | 8 | routes | routes | route/api, import-state, filesystem |
+| `tests/test_email_owner_scope.py` | 474 | 9 | 9 | 0 | security | owner_scope | route/api, db/session, filesystem, async/threading |
+| `tests/test_service_health.py` | 472 | 47 | 42 | 0 | uncategorized | service_health | async/threading |
+| `tests/test_agent_loop.py` | 469 | 52 | 52 | 5 | uncategorized | agent_loop | db/session, import-state |
+| `tests/test_kv_cache_invalidation_2927.py` | 463 | 8 | 8 | 0 | uncategorized | kv_cache_invalidation_2927 | route/api, db/session, import-state, async/threading |
+| `tests/test_helpers_import_state.py` | 426 | 26 | 26 | 0 | helpers | helpers | route/api, db/session, import-state |
+| `tests/test_endpoint_owner_scope_followup.py` | 414 | 11 | 11 | 0 | security | owner_scope | route/api, db/session, filesystem |
+| `tests/test_endpoint_probing.py` | 411 | 34 | 30 | 6 | uncategorized | endpoint_probing | route/api, db/session, import-state |
+| `tests/test_imap_leak_fixes.py` | 404 | 15 | 15 | 0 | uncategorized | imap_leak_fixes | route/api, db/session, security, filesystem |
+| `tests/test_companion_readonly.py` | 402 | 17 | 17 | 0 | uncategorized | companion_readonly | db/session, import-state |
+| `tests/test_api_chat_security.py` | 401 | 22 | 8 | 0 | security | security | route/api, db/session, import-state, filesystem, async/threading |
+| `tests/test_upload_handler_atomicity.py` | 401 | 9 | 9 | 0 | uncategorized | upload_handler_atomicity | filesystem, async/threading |
+| `tests/test_run_focus.py` | 399 | 47 | 44 | 0 | uncategorized | run_focus | security, filesystem, subprocess/script, ui/static |
+| `tests/test_auth_regressions.py` | 375 | 15 | 15 | 0 | security | auth | route/api, db/session, import-state, async/threading |
+| `tests/test_calendar_owner_scope.py` | 345 | 7 | 7 | 0 | security | owner_scope | route/api, db/session, import-state, filesystem, async/threading, ui/static |
+| `tests/test_null_owner_gates.py` | 342 | 20 | 20 | 0 | security | owner | route/api, db/session, import-state |
+| `tests/test_agent_migration_manifest.py` | 340 | 15 | 15 | 0 | uncategorized | agent_migration_manifest | import-state, filesystem |
+| `tests/test_calendar_recurrence.py` | 338 | 19 | 19 | 0 | services | calendar | - |
+| `tests/test_tool_policy.py` | 330 | 13 | 13 | 0 | uncategorized | tool_policy | import-state, async/threading |
+| `tests/test_workspace_confine.py` | 328 | 18 | 18 | 0 | uncategorized | workspace_confine | route/api, filesystem, subprocess/script, async/threading |
+| `tests/test_diffusion_server_security.py` | 325 | 14 | 14 | 0 | security | security | route/api, import-state, security, filesystem, async/threading, ui/static |
+
+## Split planning candidates
+
+This section is generated from metrics, not from manual judgement.
+Files are included when they meet at least one threshold:
+
+- at least 300 physical lines; or
+- at least 20 collected pytest items.
+
+These are planning candidates only. A later split PR still needs a focused manual review of each file before moving tests.
+
+| File | Why included | Setup/risk signals | Suggested handling |
+|---|---|---|---|
+| `tests/test_model_routes.py` | 1778 lines, 139 collected tests | route/api, db/session, import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_security_regressions.py` | 1224 lines, 92 collected tests | route/api, db/session, import-state, security, filesystem, async/threading, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_provider_classification.py` | 67 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_cookbook_helpers.py` | 912 lines, 65 collected tests | route/api, filesystem, subprocess/script, async/threading, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_shell_routes.py` | 481 lines, 63 collected tests | route/api, import-state, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_pr_blocker_audit.py` | 964 lines, 58 collected tests | import-state, security, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_provider_endpoints.py` | 58 collected tests | subprocess/script | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_agent_loop.py` | 469 lines, 52 collected tests | db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_service_health.py` | 472 lines, 47 collected tests | async/threading | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_run_focus.py` | 399 lines, 47 collected tests | security, filesystem, subprocess/script, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_llm_core_temperature.py` | 41 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_endpoint_probing.py` | 411 lines, 34 collected tests | route/api, db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_llm_core_anthropic_temp_omit.py` | 32 collected tests | db/session | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_chat_helpers.py` | 31 collected tests | route/api | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_provider_detection.py` | 31 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_model_context.py` | 30 collected tests | db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_endpoint_resolver.py` | 30 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_embedding_lanes.py` | 1104 lines, 29 collected tests | filesystem | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_upload_limits_centralized.py` | 29 collected tests | import-state, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_email_oauth.py` | 580 lines, 28 collected tests | route/api, db/session, security, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_review_regressions.py` | 930 lines, 26 collected tests | route/api, db/session, import-state, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_rename_user_owner_sync.py` | 686 lines, 26 collected tests | route/api, db/session, import-state, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_helpers_import_state.py` | 426 lines, 26 collected tests | route/api, db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_taxonomy.py` | 26 collected tests | security, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_tool_path_confinement.py` | 24 collected tests | import-state, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_copilot.py` | 23 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_research_utils.py` | 23 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_api_chat_security.py` | 401 lines, 22 collected tests | route/api, db/session, import-state, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_tool_support_heuristic.py` | 22 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_platform_compat.py` | 318 lines, 21 collected tests | import-state, filesystem, subprocess/script | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_context_compactor.py` | 21 collected tests | db/session, import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_prompt_security.py` | 21 collected tests | No obvious setup signals from static scan. | Good first manual-review candidate if test themes are cohesive. |
+| `tests/test_null_owner_gates.py` | 342 lines, 20 collected tests | route/api, db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_youtube_handler_consolidation.py` | 20 collected tests | route/api, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_calendar_recurrence.py` | 338 lines | No obvious setup signals from static scan. | Plan split boundaries before editing. |
+| `tests/test_workspace_confine.py` | 328 lines | route/api, filesystem, subprocess/script, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_api_token_routes.py` | 578 lines | route/api, db/session, import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_companion_readonly.py` | 402 lines | db/session, import-state | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_set_admin.py` | 317 lines | route/api, import-state, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_imap_leak_fixes.py` | 404 lines | route/api, db/session, security, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_auth_regressions.py` | 375 lines | route/api, db/session, import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_agent_migration_manifest.py` | 340 lines | import-state, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_diffusion_server_security.py` | 325 lines | route/api, import-state, security, filesystem, async/threading, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_tool_policy.py` | 330 lines | import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_endpoint_owner_scope_followup.py` | 414 lines | route/api, db/session, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_upload_routes_owner_scope.py` | 315 lines | route/api, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_email_owner_scope.py` | 474 lines | route/api, db/session, filesystem, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_upload_handler_atomicity.py` | 401 lines | filesystem, async/threading | Plan split boundaries before editing. |
+| `tests/test_kv_cache_invalidation_2927.py` | 463 lines | route/api, db/session, import-state, async/threading | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_calendar_owner_scope.py` | 345 lines | route/api, db/session, import-state, filesystem, async/threading, ui/static | Defer mechanical split until setup/risk boundaries are mapped. |
+| `tests/test_skills_manager_owner_isolation.py` | 306 lines | import-state, filesystem | Defer mechanical split until setup/risk boundaries are mapped. |
+
+## Taxonomy coverage gaps among split candidates
+
+`uncategorized` is a current taxonomy area, not a builder failure.
+This plan does not reclassify tests because taxonomy changes should be reviewed separately from oversized-file split planning.
+
+Before using any of these files as a split target, first decide whether the taxonomy should be refined in a separate focused issue/PR.
+
+| File | Lines | Collected tests | Sub-area | Signals | Suggested follow-up |
+|---|---:|---:|---|---|---|
+| `tests/test_pr_blocker_audit.py` | 964 | 58 | pr_blocker_audit | import-state, security, filesystem | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_agent_loop.py` | 469 | 52 | agent_loop | db/session, import-state | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_service_health.py` | 472 | 47 | service_health | async/threading | Review taxonomy mapping before using as a split target. |
+| `tests/test_run_focus.py` | 399 | 47 | run_focus | security, filesystem, subprocess/script, ui/static | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_endpoint_probing.py` | 411 | 34 | endpoint_probing | route/api, db/session, import-state | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_chat_helpers.py` | 264 | 31 | chat_helpers | route/api | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_model_context.py` | 251 | 30 | model_context | db/session, import-state | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_endpoint_resolver.py` | 148 | 30 | endpoint_resolver | - | Review taxonomy mapping before using as a split target. |
+| `tests/test_upload_limits_centralized.py` | 110 | 29 | upload_limits_centralized | import-state, filesystem | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_review_regressions.py` | 930 | 26 | review_regressions | route/api, db/session, import-state, filesystem, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_taxonomy.py` | 145 | 26 | taxonomy | security, ui/static | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_copilot.py` | 170 | 23 | copilot | - | Review taxonomy mapping before using as a split target. |
+| `tests/test_tool_support_heuristic.py` | 166 | 22 | tool_support_heuristic | - | Review taxonomy mapping before using as a split target. |
+| `tests/test_platform_compat.py` | 318 | 21 | platform_compat | import-state, filesystem, subprocess/script | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_context_compactor.py` | 233 | 21 | context_compactor | db/session, import-state, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_youtube_handler_consolidation.py` | 104 | 20 | youtube_handler_consolidation | route/api, import-state | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_workspace_confine.py` | 328 | 18 | workspace_confine | route/api, filesystem, subprocess/script, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_companion_readonly.py` | 402 | 17 | companion_readonly | db/session, import-state | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_set_admin.py` | 317 | 17 | set_admin | route/api, import-state, filesystem, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_imap_leak_fixes.py` | 404 | 15 | imap_leak_fixes | route/api, db/session, security, filesystem | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_agent_migration_manifest.py` | 340 | 15 | agent_migration_manifest | import-state, filesystem | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_tool_policy.py` | 330 | 13 | tool_policy | import-state, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+| `tests/test_upload_handler_atomicity.py` | 401 | 9 | upload_handler_atomicity | filesystem, async/threading | Review taxonomy mapping before using as a split target. |
+| `tests/test_kv_cache_invalidation_2927.py` | 463 | 8 | kv_cache_invalidation_2927 | route/api, db/session, import-state, async/threading | Review taxonomy and setup/risk boundaries before any split. |
+
+## Suggested first manual-review candidates
+
+These are not automatic split approvals. They are categorized candidates with enough size/collection value and no route/API, DB/session, import-state, or security signal from the static scan.
+
+Files still in the `uncategorized` taxonomy area are listed separately below so taxonomy review does not get mixed into the first split decision.
+
+| File | Lines | Collected tests | Area | Sub-area | Signals | Why this is a candidate |
+|---|---:|---:|---|---|---|---|
+| `tests/test_provider_classification.py` | 188 | 67 | services | provider | - | 67 collected tests |
+| `tests/test_provider_endpoints.py` | 241 | 58 | services | provider | subprocess/script | 58 collected tests |
+| `tests/test_llm_core_temperature.py` | 196 | 41 | services | llm | - | 41 collected tests |
+| `tests/test_provider_detection.py` | 148 | 31 | services | provider | - | 31 collected tests |
+| `tests/test_embedding_lanes.py` | 1104 | 29 | services | embedding | filesystem | 1104 lines, 29 collected tests |
+| `tests/test_research_utils.py` | 97 | 23 | services | research | - | 23 collected tests |
+| `tests/test_prompt_security.py` | 203 | 21 | security | security | - | 21 collected tests |
+| `tests/test_calendar_recurrence.py` | 338 | 19 | services | calendar | - | 338 lines |
+
+## High-risk candidates to defer first
+
+These files may still be split later, but not as the first implementation slice without a separate manual boundary review.
+
+| File | Lines | Collected tests | High-risk signals |
+|---|---:|---:|---|
+| `tests/test_model_routes.py` | 1778 | 139 | db/session, import-state, route/api |
+| `tests/test_security_regressions.py` | 1224 | 92 | db/session, import-state, route/api, security |
+| `tests/test_cookbook_helpers.py` | 912 | 65 | route/api |
+| `tests/test_shell_routes.py` | 481 | 63 | import-state, route/api |
+| `tests/test_pr_blocker_audit.py` | 964 | 58 | import-state, security |
+| `tests/test_agent_loop.py` | 469 | 52 | db/session, import-state |
+| `tests/test_run_focus.py` | 399 | 47 | security |
+| `tests/test_endpoint_probing.py` | 411 | 34 | db/session, import-state, route/api |
+| `tests/test_llm_core_anthropic_temp_omit.py` | 94 | 32 | db/session |
+| `tests/test_chat_helpers.py` | 264 | 31 | route/api |
+| `tests/test_model_context.py` | 251 | 30 | db/session, import-state |
+| `tests/test_upload_limits_centralized.py` | 110 | 29 | import-state |
+| `tests/test_email_oauth.py` | 580 | 28 | db/session, route/api, security |
+| `tests/test_review_regressions.py` | 930 | 26 | db/session, import-state, route/api |
+| `tests/test_rename_user_owner_sync.py` | 686 | 26 | db/session, import-state, route/api |
+
+## Rules for future split PRs
+
+- One file or one coherent file-family per PR.
+- No assertion rewrites mixed with file moves.
+- No helper extraction mixed with file moves.
+- No production code changes.
+- No CI workflow changes.
+- Preserve existing markers and taxonomy unless the split issue explicitly says otherwise.
+- Validate the original file's collected tests before and after the split.
+- Validate any neighboring taxonomy/focused-runner behavior if paths change.
+- Treat files with route/API, DB/session, import-state, or security signals as higher-risk until manually reviewed.
+
+## Suggested next step
+
+Use this plan to choose the first actual oversized-file split issue.
+The first split should prefer a file with high review value and low setup risk.
+Do not start a split PR from this planning issue alone if the file's boundaries are still ambiguous.
+
+## Reproduction command
+
+This document was generated with:
+
+```bash
+.venv/bin/python tests/tools/build_oversized_test_split_plan.py
+```
+
+## Freshness check
+
+After editing the builder or rebasing the branch, regenerate the plan and confirm no unexpected plan drift:
+
+```bash
+.venv/bin/python tests/tools/build_oversized_test_split_plan.py
+git diff --exit-code -- tests/OVERSIZED_TEST_SPLIT_PLAN.md
+```

tests/test_api_chat_security.py

@@ -219,9 +219,6 @@ class _WebhookManager:
     async def fire(self, event, payload):
         return None
 
-    def fire_and_forget(self, event, payload):
-        return None
-
 
 def _install_sync_chat_stubs(monkeypatch):
     # FastAPI checks for python_multipart at import time when Form is used;

tests/test_checkin_digest_owner_scope.py

@@ -1,70 +0,0 @@
-"""Check-in calendar digest must be scoped to the task owner.
-
-The digest query selected CalendarEvent with no owner scope, so a scheduled
-check-in for one user pulled EVERY user's calendar events (summaries,
-locations) into their digest — a cross-tenant leak. Ownership lives on
-CalendarCal.owner; the query must join it, like routes/calendar_routes.
-"""
-import tempfile
-import uuid
-from datetime import datetime
-
-import pytest
-from sqlalchemy import create_engine
-from sqlalchemy.orm import sessionmaker
-from sqlalchemy.pool import NullPool
-
-import core.database as cdb
-from core.database import CalendarEvent, CalendarCal
-from src.task_scheduler import _checkin_calendar_events
-
-_TMPDB = tempfile.NamedTemporaryFile(suffix=".db", delete=False)
-_ENGINE = create_engine(f"sqlite:///{_TMPDB.name}", connect_args={"check_same_thread": False}, poolclass=NullPool)
-cdb.Base.metadata.create_all(_ENGINE)
-_TS = sessionmaker(bind=_ENGINE, autoflush=False, autocommit=False)
-
-
-def _seed():
-    db = _TS()
-    try:
-        db.query(CalendarEvent).delete(); db.query(CalendarCal).delete()
-        db.add(CalendarCal(id="calA", owner="alice", name="A"))
-        db.add(CalendarCal(id="calB", owner="bob", name="B"))
-        db.add(CalendarEvent(uid="a1", calendar_id="calA", summary="Alice mtg",
-                             dtstart=datetime(2026, 6, 10, 9, 0),
-                             dtend=datetime(2026, 6, 10, 10, 0), status="confirmed"))
-        db.add(CalendarEvent(uid="b1", calendar_id="calB", summary="Bob secret",
-                             dtstart=datetime(2026, 6, 10, 10, 0),
-                             dtend=datetime(2026, 6, 10, 11, 0), status="confirmed"))
-        db.commit()
-    finally:
-        db.close()
-
-
-def test_digest_only_returns_owner_events():
-    _seed()
-    db = _TS()
-    try:
-        s, e = datetime(2026, 6, 1), datetime(2026, 6, 30)
-        alice = _checkin_calendar_events(db, "alice", s, e)
-        assert [ev.summary for ev in alice] == ["Alice mtg"]  # not Bob's
-        bob = _checkin_calendar_events(db, "bob", s, e)
-        assert [ev.summary for ev in bob] == ["Bob secret"]
-    finally:
-        db.close()
-
-
-def test_cancelled_excluded_and_window_respected():
-    _seed()
-    db = _TS()
-    try:
-        db2 = _TS()
-        db2.add(CalendarEvent(uid="a2", calendar_id="calA", summary="cancelled",
-                              dtstart=datetime(2026, 6, 11),
-                              dtend=datetime(2026, 6, 11, 1, 0), status="cancelled"))
-        db2.commit(); db2.close()
-        s, e = datetime(2026, 6, 1), datetime(2026, 6, 30)
-        out = _checkin_calendar_events(db, "alice", s, e)
-        assert "cancelled" not in [ev.summary for ev in out]
-    finally:
-        db.close()

tests/test_fastembed_cache_path.py

@@ -1,69 +0,0 @@
-"""Regression: FASTEMBED_CACHE_DIR must tolerate a PRESENT-but-EMPTY
-FASTEMBED_CACHE_PATH.
-
-docker-compose.yml injects ``FASTEMBED_CACHE_PATH=${FASTEMBED_CACHE_PATH:-}``,
-which sets the variable to ``""`` when the host has not defined it. The old
-``os.getenv("FASTEMBED_CACHE_PATH", default)`` only used the default when the
-variable was ABSENT, so an empty value made ``FASTEMBED_CACHE_DIR == ""`` →
-``os.makedirs("")`` raised ``[Errno 2] No such file or directory: ''`` →
-FastEmbed failed to initialise and every vector feature (RAG, semantic memory,
-tool index) silently degraded on the default Docker stack.
-
-These tests pin the fix: empty is treated like absent → use the DATA_DIR
-default, while an explicit non-empty override is still honoured.
-"""
-
-from __future__ import annotations
-
-import importlib
-import os
-
-import src.constants as constants
-
-
-def _reload_with(monkeypatch, value):
-    """Reload src.constants with FASTEMBED_CACHE_PATH set to ``value`` (or
-    removed when ``value`` is None) and return the reloaded module."""
-    if value is None:
-        monkeypatch.delenv("FASTEMBED_CACHE_PATH", raising=False)
-    else:
-        monkeypatch.setenv("FASTEMBED_CACHE_PATH", value)
-    return importlib.reload(constants)
-
-
-def _restore(monkeypatch):
-    """Return the module to its env-default state so reloading it here does
-    not leak a test-specific FASTEMBED_CACHE_DIR into other tests."""
-    monkeypatch.delenv("FASTEMBED_CACHE_PATH", raising=False)
-    importlib.reload(constants)
-
-
-def test_empty_fastembed_cache_path_falls_back_to_default(monkeypatch):
-    """The bug: an empty FASTEMBED_CACHE_PATH (exactly what Docker injects)
-    must fall back to the DATA_DIR default, never the empty string."""
-    try:
-        mod = _reload_with(monkeypatch, "")
-        assert mod.FASTEMBED_CACHE_DIR, "empty env must not yield an empty path"
-        assert mod.FASTEMBED_CACHE_DIR == os.path.join(mod.DATA_DIR, "fastembed_cache")
-    finally:
-        _restore(monkeypatch)
-
-
-def test_unset_fastembed_cache_path_uses_default(monkeypatch):
-    """Sanity: an absent variable also resolves to the default."""
-    try:
-        mod = _reload_with(monkeypatch, None)
-        assert mod.FASTEMBED_CACHE_DIR == os.path.join(mod.DATA_DIR, "fastembed_cache")
-    finally:
-        _restore(monkeypatch)
-
-
-def test_explicit_fastembed_cache_path_is_respected(monkeypatch):
-    """A real explicit override must still win — the fix only changes the
-    empty-value handling, not the documented FASTEMBED_CACHE_PATH override."""
-    custom = os.path.join("custom", "fastembed-cache")
-    try:
-        mod = _reload_with(monkeypatch, custom)
-        assert mod.FASTEMBED_CACHE_DIR == custom
-    finally:
-        _restore(monkeypatch)

tests/test_mcp_memory_owner_scope.py

@@ -1,150 +0,0 @@
-import asyncio
-
-import mcp_servers.memory_server as memory_server
-from src.memory import MemoryManager
-
-
-class FakeVector:
-    healthy = True
-
-    def __init__(self):
-        self.added = []
-        self.removed = []
-
-    def add(self, memory_id, text):
-        self.added.append((memory_id, text))
-
-    def remove(self, memory_id):
-        self.removed.append(memory_id)
-
-
-def _tool_text(arguments):
-    result = asyncio.run(memory_server.call_tool("manage_memory", arguments))
-    return result[0].text
-
-
-def _entry(manager, text, owner=None, memory_id=None, category="fact"):
-    entry = manager.add_entry(text, owner=owner, category=category)
-    if memory_id:
-        entry["id"] = memory_id
-    return entry
-
-
-def _configure_server(monkeypatch, manager, vector=None):
-    monkeypatch.setattr(memory_server, "_memory_manager", manager)
-    monkeypatch.setattr(memory_server, "_memory_vector", vector)
-    monkeypatch.setattr(memory_server, "_initialized", True)
-    for key in memory_server._OWNER_ENV_KEYS:
-        monkeypatch.delenv(key, raising=False)
-
-
-def test_mcp_memory_uses_configured_owner_for_all_operations(monkeypatch, tmp_path):
-    manager = MemoryManager(str(tmp_path))
-    vector = FakeVector()
-    alice = _entry(
-        manager,
-        "Alice likes green tea",
-        owner="alice",
-        memory_id="aaaaaaaa-0000-0000-0000-000000000000",
-    )
-    bob = _entry(
-        manager,
-        "Bob likes espresso",
-        owner="bob",
-        memory_id="bbbbbbbb-0000-0000-0000-000000000000",
-    )
-    manager.save([alice, bob])
-    _configure_server(monkeypatch, manager, vector)
-    monkeypatch.setenv("ODYSSEUS_MCP_MEMORY_OWNER", "alice")
-
-    list_text = _tool_text({"action": "list"})
-    assert "Alice likes green tea" in list_text
-    assert "Bob likes espresso" not in list_text
-
-    search_text = _tool_text({"action": "search", "text": "likes"})
-    assert "Alice likes green tea" in search_text
-    assert "Bob likes espresso" not in search_text
-
-    add_text = _tool_text({
-        "action": "add",
-        "text": "Alice prefers concise notes",
-        "category": "preference",
-    })
-    assert "Memory added" in add_text
-    added = next(
-        entry for entry in manager.load_all()
-        if entry["text"] == "Alice prefers concise notes"
-    )
-    assert added["owner"] == "alice"
-    assert vector.added == [(added["id"], "Alice prefers concise notes")]
-
-    edit_text = _tool_text({
-        "action": "edit",
-        "memory_id": bob["id"][:8],
-        "text": "Bob changed",
-    })
-    assert edit_text == "Error: Memory 'bbbbbbbb' not found"
-    bob_after_edit = next(
-        entry for entry in manager.load_all()
-        if entry["id"] == bob["id"]
-    )
-    assert bob_after_edit["text"] == "Bob likes espresso"
-
-    delete_text = _tool_text({"action": "delete", "memory_id": bob["id"][:8]})
-    assert delete_text == "Error: Memory 'bbbbbbbb' not found"
-    assert any(entry["id"] == bob["id"] for entry in manager.load_all())
-
-
-def test_mcp_memory_fails_closed_without_owner_for_owner_scoped_store(monkeypatch, tmp_path):
-    manager = MemoryManager(str(tmp_path))
-    alice = _entry(manager, "Alice private memory", owner="alice", memory_id="aaaaaaaa-0000")
-    bob = _entry(manager, "Bob private memory", owner="bob", memory_id="bbbbbbbb-0000")
-    manager.save([alice, bob])
-    _configure_server(monkeypatch, manager, FakeVector())
-    before = manager.load_all()
-
-    actions = [
-        {"action": "list"},
-        {"action": "search", "text": "private"},
-        {"action": "add", "text": "new ownerless memory"},
-        {"action": "edit", "memory_id": alice["id"][:8], "text": "changed"},
-        {"action": "delete", "memory_id": alice["id"][:8]},
-    ]
-
-    for arguments in actions:
-        assert _tool_text(arguments).startswith("Error: Memory MCP owner is not configured")
-
-    assert manager.load_all() == before
-
-
-def test_mcp_memory_preserves_ownerless_local_behavior(monkeypatch, tmp_path):
-    manager = MemoryManager(str(tmp_path))
-    legacy = _entry(
-        manager,
-        "Legacy local memory",
-        memory_id="llllllll-0000-0000-0000-000000000000",
-    )
-    manager.save([legacy])
-    _configure_server(monkeypatch, manager, FakeVector())
-
-    assert "Legacy local memory" in _tool_text({"action": "list"})
-    assert "Legacy local memory" in _tool_text({"action": "search", "text": "legacy"})
-
-    add_text = _tool_text({"action": "add", "text": "Another local memory"})
-    assert "Memory added" in add_text
-    added = next(
-        entry for entry in manager.load_all()
-        if entry["text"] == "Another local memory"
-    )
-    assert "owner" not in added
-
-    assert _tool_text({
-        "action": "edit",
-        "memory_id": legacy["id"][:8],
-        "text": "Updated local memory",
-    }) == "Memory updated: Updated local memory"
-    assert any(entry["text"] == "Updated local memory" for entry in manager.load_all())
-
-    delete_text = _tool_text({"action": "delete", "memory_id": legacy["id"][:8]})
-    assert delete_text.startswith("Memory deleted:")
-    assert all(entry["id"] != legacy["id"] for entry in manager.load_all())

tests/test_web_user_agent_constant.py

@@ -1,18 +0,0 @@
-"""The web scraping path routes its User-Agent through one constant.
-
-Guards the dedup: web_fetch / web_search outbound UAs go through
-WEB_FETCH_USER_AGENT, so a stale or bare Mozilla string cannot be re-inlined in
-the search sources.
-"""
-from pathlib import Path
-
-_SEARCH = Path(__file__).resolve().parent.parent / "services" / "search"
-
-
-def test_search_sources_have_no_inline_mozilla_ua():
-    offenders = [
-        str(py.relative_to(_SEARCH.parent.parent))
-        for py in _SEARCH.rglob("*.py")
-        if "Mozilla/" in py.read_text(encoding="utf-8")
-    ]
-    assert not offenders, f"inline Mozilla UA found; use WEB_FETCH_USER_AGENT: {offenders}"

tests/tools/build_oversized_test_split_plan.py

@@ -0,0 +1,574 @@
+#!/usr/bin/env python3
+"""Build the oversized test-file split plan for issue #3983.
+
+The output is a planning document only. It does not move tests, rewrite
+assertions, extract helpers, or change CI.
+"""
+from __future__ import annotations
+
+import ast
+import json
+import os
+import re
+import subprocess
+import sys
+from collections import Counter
+from dataclasses import dataclass
+from pathlib import Path
+
+ROOT = Path(__file__).resolve().parents[2]
+TESTS_DIR = ROOT / "tests"
+OUTPUT = TESTS_DIR / "OVERSIZED_TEST_SPLIT_PLAN.md"
+RAW_OUTPUT = Path("/tmp/oversized-test-file-metrics.json")
+
+LARGE_LINE_THRESHOLD = 300
+LARGE_NODE_THRESHOLD = 20
+TOP_LIMIT = 30
+
+HIGH_RISK_SIGNALS = {"route/api", "db/session", "import-state", "security"}
+
+
+@dataclass(frozen=True)
+class FileMetric:
+    path: str
+    lines: int
+    nonblank: int
+    test_defs: int
+    test_classes: int
+    collected: int
+    area: str
+    sub_area: str
+    signals: tuple[str, ...]
+
+
+def read_text(path: Path) -> str:
+    return path.read_text(encoding="utf-8", errors="replace")
+
+
+def count_ast_tests(text: str) -> tuple[int, int]:
+    tree = ast.parse(text)
+    test_defs = 0
+    test_classes = 0
+
+    for node in ast.walk(tree):
+        if isinstance(node, (ast.FunctionDef, ast.AsyncFunctionDef)):
+            if node.name.startswith("test_"):
+                test_defs += 1
+        elif isinstance(node, ast.ClassDef):
+            if node.name.startswith("Test"):
+                test_classes += 1
+
+    return test_defs, test_classes
+
+
+def load_taxonomy_classifier():
+    sys.path.insert(0, str(ROOT))
+    from tests._taxonomy import classify_test_path
+
+    return classify_test_path
+
+
+def classify(path: Path, classify_test_path) -> tuple[str, str]:
+    rel_path = Path(path.relative_to(ROOT).as_posix())
+
+    try:
+        result = classify_test_path(rel_path)
+    except Exception:
+        return "unknown", "unknown"
+
+    return getattr(result, "area", "unknown"), getattr(result, "sub_area", "unknown")
+
+
+def collect_node_counts() -> Counter[str]:
+    cmd = [
+        sys.executable,
+        "-m",
+        "pytest",
+        "--collect-only",
+        "-q",
+        "tests",
+    ]
+    env = dict(os.environ)
+    env["PY_COLORS"] = "0"
+
+    result = subprocess.run(
+        cmd,
+        cwd=ROOT,
+        env=env,
+        text=True,
+        capture_output=True,
+    )
+
+    if result.returncode != 0:
+        print(result.stdout)
+        print(result.stderr, file=sys.stderr)
+        raise SystemExit(result.returncode)
+
+    counts: Counter[str] = Counter()
+    for line in result.stdout.splitlines():
+        line = line.strip()
+        if "::" not in line:
+            continue
+        if not line.startswith("tests/"):
+            continue
+        file_path = line.split("::", 1)[0]
+        counts[file_path] += 1
+
+    return counts
+
+
+def detect_signals(text: str, path: str) -> tuple[str, ...]:
+    signal_patterns = {
+        "route/api": [
+            r"\bTestClient\b",
+            r"\bapp\.",
+            r"\broutes\.",
+            r"\bfrom routes\b",
+            r"\bimport routes\b",
+        ],
+        "db/session": [
+            r"\bSessionLocal\b",
+            r"\bsqlite\b",
+            r"\bDATABASE_URL\b",
+            r"\bcore\.database\b",
+            r"\bdb\.query\b",
+            r"\bcommit\(",
+        ],
+        "import-state": [
+            r"\bsys\.modules\b",
+            r"\bimportlib\b",
+            r"\bclear_module\b",
+            r"\bpreserve_import_state\b",
+            r"\bmonkeypatch\.setitem\b",
+        ],
+        "security": [
+            r"\bsecurity\b",
+            r"\bssrf\b",
+            r"\bpath traversal\b",
+            r"\bcsrf\b",
+            r"\bpermission\b",
+        ],
+        "filesystem": [
+            r"\btmp_path\b",
+            r"\bTemporaryDirectory\b",
+            r"\bPath\(",
+            r"\bmkdir\b",
+            r"\bwrite_text\b",
+            r"\bread_text\b",
+        ],
+        "subprocess/script": [
+            r"\bsubprocess\b",
+            r"\brunpy\b",
+            r"\bload_script\b",
+            r"\bsys\.argv\b",
+        ],
+        "async/threading": [
+            r"\basyncio\b",
+            r"\bthreading\b",
+            r"\bconcurrent\.futures\b",
+            r"\bThreadPoolExecutor\b",
+        ],
+        "ui/static": [
+            r"\bstatic/",
+            r"\bjsdom\b",
+            r"\bnode\b",
+            r"\.js\b",
+        ],
+    }
+
+    signals = []
+    for name, patterns in signal_patterns.items():
+        if any(re.search(pattern, text, flags=re.IGNORECASE) for pattern in patterns):
+            signals.append(name)
+
+    if path.startswith("tests/cli/"):
+        signals.append("cli-directory")
+
+    return tuple(signals)
+
+
+def metric_for(path: Path, node_counts: Counter[str], classify_test_path) -> FileMetric:
+    rel = path.relative_to(ROOT).as_posix()
+    text = read_text(path)
+    lines = len(text.splitlines())
+    nonblank = sum(1 for line in text.splitlines() if line.strip())
+    test_defs, test_classes = count_ast_tests(text)
+    area, sub_area = classify(path, classify_test_path)
+
+    return FileMetric(
+        path=rel,
+        lines=lines,
+        nonblank=nonblank,
+        test_defs=test_defs,
+        test_classes=test_classes,
+        collected=node_counts.get(rel, 0),
+        area=area,
+        sub_area=sub_area,
+        signals=detect_signals(text, rel),
+    )
+
+
+def test_files() -> list[Path]:
+    return sorted(TESTS_DIR.rglob("test_*.py"))
+
+
+def as_metric_row(metric: FileMetric) -> str:
+    signals = ", ".join(metric.signals) if metric.signals else "-"
+    return (
+        f"| `{metric.path}` | {metric.lines} | {metric.collected} | "
+        f"{metric.test_defs} | {metric.test_classes} | "
+        f"{metric.area} | {metric.sub_area} | {signals} |"
+    )
+
+
+def metric_table(title: str, metrics: list[FileMetric]) -> list[str]:
+    lines = [
+        f"## {title}",
+        "",
+        "| File | Lines | Collected tests | Test defs | Test classes | Area | Sub-area | Signals |",
+        "|---|---:|---:|---:|---:|---|---|---|",
+    ]
+    lines.extend(as_metric_row(metric) for metric in metrics)
+    lines.append("")
+    return lines
+
+
+def candidate_metrics(metrics: list[FileMetric]) -> list[FileMetric]:
+    return [
+        metric
+        for metric in metrics
+        if metric.lines >= LARGE_LINE_THRESHOLD
+        or metric.collected >= LARGE_NODE_THRESHOLD
+    ]
+
+
+def include_reasons(metric: FileMetric) -> str:
+    reasons = []
+    if metric.lines >= LARGE_LINE_THRESHOLD:
+        reasons.append(f"{metric.lines} lines")
+    if metric.collected >= LARGE_NODE_THRESHOLD:
+        reasons.append(f"{metric.collected} collected tests")
+    return ", ".join(reasons)
+
+
+def risk_notes(metric: FileMetric) -> str:
+    if not metric.signals:
+        return "No obvious setup signals from static scan."
+    return ", ".join(metric.signals)
+
+
+def suggested_handling(metric: FileMetric) -> str:
+    if HIGH_RISK_SIGNALS.intersection(metric.signals):
+        return "Defer mechanical split until setup/risk boundaries are mapped."
+    if metric.collected >= LARGE_NODE_THRESHOLD:
+        return "Good first manual-review candidate if test themes are cohesive."
+    return "Plan split boundaries before editing."
+
+
+def candidate_section(metrics: list[FileMetric]) -> list[str]:
+    lines = [
+        "## Split planning candidates",
+        "",
+        "This section is generated from metrics, not from manual judgement.",
+        "Files are included when they meet at least one threshold:",
+        "",
+        f"- at least {LARGE_LINE_THRESHOLD} physical lines; or",
+        f"- at least {LARGE_NODE_THRESHOLD} collected pytest items.",
+        "",
+        "These are planning candidates only. A later split PR still needs a focused manual review of each file before moving tests.",
+        "",
+        "| File | Why included | Setup/risk signals | Suggested handling |",
+        "|---|---|---|---|",
+    ]
+
+    for metric in metrics:
+        lines.append(
+            f"| `{metric.path}` | {include_reasons(metric)} | "
+            f"{risk_notes(metric)} | {suggested_handling(metric)} |"
+        )
+
+    lines.append("")
+    return lines
+
+
+def first_manual_review_section(metrics: list[FileMetric]) -> list[str]:
+    low_risk = [
+        metric
+        for metric in metrics
+        if metric.area != "uncategorized"
+        and not HIGH_RISK_SIGNALS.intersection(metric.signals)
+    ]
+    low_risk = sorted(low_risk, key=lambda m: (m.collected, m.lines), reverse=True)
+
+    lines = [
+        "## Suggested first manual-review candidates",
+        "",
+        "These are not automatic split approvals. They are categorized candidates with enough size/collection value and no route/API, DB/session, import-state, or security signal from the static scan.",
+        "",
+        "Files still in the `uncategorized` taxonomy area are listed separately below so taxonomy review does not get mixed into the first split decision.",
+        "",
+        "| File | Lines | Collected tests | Area | Sub-area | Signals | Why this is a candidate |",
+        "|---|---:|---:|---|---|---|---|",
+    ]
+
+    if not low_risk:
+        lines.append("| _None_ | - | - | - | - | - | - |")
+
+    for metric in low_risk[:10]:
+        signals = ", ".join(metric.signals) if metric.signals else "-"
+        lines.append(
+            f"| `{metric.path}` | {metric.lines} | {metric.collected} | "
+            f"{metric.area} | {metric.sub_area} | {signals} | {include_reasons(metric)} |"
+        )
+
+    lines.append("")
+    return lines
+
+
+def taxonomy_gap_section(metrics: list[FileMetric]) -> list[str]:
+    uncategorized = [
+        metric
+        for metric in metrics
+        if metric.area == "uncategorized"
+    ]
+    uncategorized = sorted(
+        uncategorized,
+        key=lambda m: (m.collected, m.lines),
+        reverse=True,
+    )
+
+    lines = [
+        "## Taxonomy coverage gaps among split candidates",
+        "",
+        "`uncategorized` is a current taxonomy area, not a builder failure.",
+        "This plan does not reclassify tests because taxonomy changes should be reviewed separately from oversized-file split planning.",
+        "",
+        "Before using any of these files as a split target, first decide whether the taxonomy should be refined in a separate focused issue/PR.",
+        "",
+        "| File | Lines | Collected tests | Sub-area | Signals | Suggested follow-up |",
+        "|---|---:|---:|---|---|---|",
+    ]
+
+    if not uncategorized:
+        lines.append("| _None_ | - | - | - | - | - |")
+
+    for metric in uncategorized:
+        signals = ", ".join(metric.signals) if metric.signals else "-"
+        follow_up = "Review taxonomy mapping before using as a split target."
+        if HIGH_RISK_SIGNALS.intersection(metric.signals):
+            follow_up = "Review taxonomy and setup/risk boundaries before any split."
+        lines.append(
+            f"| `{metric.path}` | {metric.lines} | {metric.collected} | "
+            f"{metric.sub_area} | {signals} | {follow_up} |"
+        )
+
+    lines.append("")
+    return lines
+
+
+def deferred_section(metrics: list[FileMetric]) -> list[str]:
+    deferred = [
+        metric
+        for metric in metrics
+        if HIGH_RISK_SIGNALS.intersection(metric.signals)
+    ]
+    deferred = sorted(deferred, key=lambda m: (m.collected, m.lines), reverse=True)
+
+    lines = [
+        "## High-risk candidates to defer first",
+        "",
+        "These files may still be split later, but not as the first implementation slice without a separate manual boundary review.",
+        "",
+        "| File | Lines | Collected tests | High-risk signals |",
+        "|---|---:|---:|---|",
+    ]
+
+    for metric in deferred[:15]:
+        signals = ", ".join(sorted(HIGH_RISK_SIGNALS.intersection(metric.signals)))
+        lines.append(
+            f"| `{metric.path}` | {metric.lines} | {metric.collected} | {signals} |"
+        )
+
+    lines.append("")
+    return lines
+
+
+def write_distribution(
+    lines: list[str],
+    title: str,
+    values: Counter[str],
+    *,
+    min_count: int = 1,
+) -> None:
+    displayed = [
+        (value, count)
+        for value, count in sorted(values.items())
+        if count >= min_count
+    ]
+    omitted_values = sum(1 for count in values.values() if count < min_count)
+    omitted_files = sum(count for count in values.values() if count < min_count)
+
+    lines.extend([
+        f"{title}:",
+        "",
+        "| Value | Files |",
+        "|---|---:|",
+    ])
+    for value, count in displayed:
+        lines.append(f"| {value} | {count} |")
+
+    if omitted_values:
+        lines.extend([
+            "",
+            f"Values below {min_count} files: {omitted_values} values covering {omitted_files} files.",
+        ])
+
+    lines.append("")
+
+
+def write_report(metrics: list[FileMetric], node_count_total: int) -> None:
+    by_lines = sorted(metrics, key=lambda m: (m.lines, m.collected), reverse=True)
+    by_collected = sorted(metrics, key=lambda m: (m.collected, m.lines), reverse=True)
+    candidates = sorted(
+        candidate_metrics(metrics),
+        key=lambda m: (m.collected, m.lines),
+        reverse=True,
+    )
+
+    areas = Counter(metric.area for metric in metrics)
+    sub_areas = Counter(metric.sub_area for metric in metrics)
+
+    lines = [
+        "# Oversized Test File Split Plan",
+        "",
+        "## Purpose",
+        "",
+        "This document plans future oversized test-file splits using current repo data.",
+        "It does not move files, rewrite assertions, extract helpers, or change CI.",
+        "",
+        "## Roadmap context",
+        "",
+        "- Issue: #3983",
+        "- Parent tracker: #2523",
+        "- Follows #3973 / #3982, the report-only order-sensitivity diagnostics slice.",
+        "",
+        "## Methodology",
+        "",
+        "Metrics were generated from the current test tree using:",
+        "",
+        "- physical line counts for every recursive `test_*.py` file under `tests/`;",
+        "- AST counts for `test_*` functions and `Test*` classes;",
+        "- one `pytest --collect-only -q tests` run to count collected items per file;",
+        "- current taxonomy classification from `tests._taxonomy.classify_test_path`; and",
+        "- static setup-signal scans for route/API, DB/session, import-state, security, filesystem, subprocess/script, async/threading, and UI/static indicators.",
+        "",
+        "Static signals are not proof of risk. They are review prompts.",
+        "Future split PRs must still inspect each file manually before editing.",
+        "",
+        "## Current summary",
+        "",
+        f"- test files scanned: {len(metrics)}",
+        f"- collected pytest items counted: {node_count_total}",
+        f"- large-file threshold: {LARGE_LINE_THRESHOLD} lines",
+        f"- large-collected threshold: {LARGE_NODE_THRESHOLD} collected items",
+        "",
+    ]
+
+    write_distribution(lines, "Area distribution", areas)
+    write_distribution(lines, "Sub-area distribution", sub_areas, min_count=2)
+
+    lines.extend(metric_table("Top files by collected pytest items", by_collected[:TOP_LIMIT]))
+    lines.extend(metric_table("Top files by physical line count", by_lines[:TOP_LIMIT]))
+    lines.extend(candidate_section(candidates))
+    lines.extend(taxonomy_gap_section(candidates))
+    lines.extend(first_manual_review_section(candidates))
+    lines.extend(deferred_section(candidates))
+
+    lines.extend([
+        "## Rules for future split PRs",
+        "",
+        "- One file or one coherent file-family per PR.",
+        "- No assertion rewrites mixed with file moves.",
+        "- No helper extraction mixed with file moves.",
+        "- No production code changes.",
+        "- No CI workflow changes.",
+        "- Preserve existing markers and taxonomy unless the split issue explicitly says otherwise.",
+        "- Validate the original file's collected tests before and after the split.",
+        "- Validate any neighboring taxonomy/focused-runner behavior if paths change.",
+        "- Treat files with route/API, DB/session, import-state, or security signals as higher-risk until manually reviewed.",
+        "",
+        "## Suggested next step",
+        "",
+        "Use this plan to choose the first actual oversized-file split issue.",
+        "The first split should prefer a file with high review value and low setup risk.",
+        "Do not start a split PR from this planning issue alone if the file's boundaries are still ambiguous.",
+        "",
+        "## Reproduction command",
+        "",
+        "This document was generated with:",
+        "",
+        "```bash",
+        ".venv/bin/python tests/tools/build_oversized_test_split_plan.py",
+        "```",
+        "",
+        "## Freshness check",
+        "",
+        "After editing the builder or rebasing the branch, regenerate the plan and confirm no unexpected plan drift:",
+        "",
+        "```bash",
+        ".venv/bin/python tests/tools/build_oversized_test_split_plan.py",
+        "git diff --exit-code -- tests/OVERSIZED_TEST_SPLIT_PLAN.md",
+        "```",
+        "",
+    ])
+
+    OUTPUT.write_text("\n".join(lines), encoding="utf-8")
+
+
+def write_raw(metrics: list[FileMetric]) -> None:
+    raw = [
+        {
+            "area": metric.area,
+            "collected": metric.collected,
+            "lines": metric.lines,
+            "nonblank": metric.nonblank,
+            "path": metric.path,
+            "signals": list(metric.signals),
+            "sub_area": metric.sub_area,
+            "test_classes": metric.test_classes,
+            "test_defs": metric.test_defs,
+        }
+        for metric in metrics
+    ]
+    RAW_OUTPUT.write_text(json.dumps(raw, indent=2, sort_keys=True), encoding="utf-8")
+
+
+def assert_taxonomy_worked(metrics: list[FileMetric]) -> None:
+    if not metrics:
+        raise SystemExit("ERROR: no test files were scanned")
+
+    unknown = sum(1 for metric in metrics if metric.area == "unknown")
+    if unknown == len(metrics):
+        raise SystemExit("ERROR: taxonomy classification returned unknown for every file")
+
+
+def main() -> int:
+    if not TESTS_DIR.exists():
+        print("ERROR: tests/ directory not found", file=sys.stderr)
+        return 1
+
+    classify_test_path = load_taxonomy_classifier()
+    node_counts = collect_node_counts()
+    metrics = [metric_for(path, node_counts, classify_test_path) for path in test_files()]
+
+    assert_taxonomy_worked(metrics)
+    write_report(metrics, sum(node_counts.values()))
+    write_raw(metrics)
+
+    print(f"Wrote {OUTPUT.relative_to(ROOT)}")
+    print(f"Wrote {RAW_OUTPUT}")
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())