coretrace-python-analyzer 0.12.0__tar.gz → 0.13.0__tar.gz

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/CHANGELOG.md

@@ -1,5 +1,15 @@
 # Changelog
 
+## 0.13.0 (2026-09-26)
+
+### Precision
+
+- The text attributes and URL parameters of a request object carry no `NOSQL`. Django, REST framework, aiohttp and tornado views read a request tainted as a whole, so `request.GET["name"]` and a URL parameter were `nosql-injection` like the body. Now `GET`, `POST`, `COOKIES`, `META`, `headers`, `query_params`, aiohttp's `query` and `match_info`, tornado's `arguments` and the other text attributes are text, and so are the URL parameters a view receives after the request. The body, the files, REST framework's `data` and aiohttp's `json()` keep the kinds of the request. The other kinds are unchanged, and so is the regression corpus. A helper function the request is passed to still reads it as a whole (#162).
+
+### Plugins
+
+- A `RequestObject(symbol, text=())` model says that the inputs from an entry point, a route registrar, a typed parameter or a source are request objects, and which of their attributes are text (#162).
+
 ## 0.12.0 (2026-09-26)
 
 ### Detection

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: coretrace-python-analyzer
-Version: 0.12.0
+Version: 0.13.0
 Summary: Static security analysis infrastructure for Python code
 Project-URL: Homepage, https://github.com/CoreTrace/coretrace-python-analyzer
 Project-URL: Repository, https://github.com/CoreTrace/coretrace-python-analyzer

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/docs/plugins.md

@@ -151,6 +151,7 @@ class BottleModels(ModelPlugin):
 | `SafeArgument(symbol, argument, values, position=None, kinds=ALL)` | A call to the sink `symbol` whose `argument` (by keyword, or at `position`) denotes one of `values` is not a sink for `kinds`, such as `yaml.load` with `Loader=yaml.SafeLoader`. Values are written as the analyzer records arguments: a symbol by its canonical name, a constant as Python writes it (`True`). Only a value given explicitly makes the call safe. |
 | `TemplateRender(symbol, position=0, keyword="template_name")` | A call rendering, with autoescaping, the template named by its argument at `position` or `keyword`, such as `render_to_string`. What it returns carries no `HTML` when the name is written in the call and the project's template escapes everything it prints (see the usage guide). |
 | `Validator(symbol, kinds=ALL, argument=0)` | A callable whose truth proves its argument safe for `kinds`, such as `re.fullmatch`. A flow guarded by it is refuted when it carries no other kind into the sink, and is a hotspot otherwise: a redirect validator does not make a shell command safe. `symbol` may be a function of the project, by its project symbol (`python.hc.accounts.views._allow_redirect`), which then counts wherever it is called, in its own module too. |
+| `RequestObject(symbol, text=())` | The inputs from `symbol`, an entry point, a route registrar, a typed parameter or a source, are request objects. A view receives the request first (after `self`), then URL parameters, which are text. Reading one of the request's `text` attributes gives text, as Django's `GET` or aiohttp's `query`; its body and files keep the request's kinds. An attribute read counts only when every source of the value is such a request. |
 | `AuthorizationGuard(symbol, label)` | A decorator or a condition restricting who reaches the code, such as `login_required`; a flow behind it is a hotspot. |
 | `Members(symbol, dynamic=None, defined=(), typed=())` | What the attributes and items of an instance of the class `symbol` give, for objects that give others by name. `typed` maps an attribute to the class of what it gives, called or not, since a call denotes its callee: pymongo's `get_database` gives a `Database`, as Flask-PyMongo's `db` property does. `dynamic`, when set, is the class of every item and of every attribute outside `defined` and `typed`: a pymongo client gives a `Database` by name (`client.shop`, `client["shop"]`), a database a `Collection`. Every path to a collection then denotes the collection class, so a sink on `Collection.find` covers them all. Without `dynamic`, an attribute keeps its path and an item its container's symbol, as for any class. |
 

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/docs/usage.md

@@ -112,7 +112,7 @@ variable is not treated that way yet.
 | `command-injection` | Shell or process execution (`os.system`, `subprocess` with a string, …). |
 | `code-injection` | The code `eval` or `exec` runs; the globals and locals passed with it are data. `dangerous-eval` reports every such call, this rule the ones attacker input reaches. |
 | `sql-injection` | A database statement (`cursor.execute`, SQLAlchemy `text`, Django `raw`, …). Parameters of a parameterised query are not statements. |
-| `nosql-injection` | A NoSQL query, such as a MongoDB filter, given attacker input that may hold a structure: a JSON body giving `{"$ne": null}` matches any value. The input that may hold a structure is JSON, the raw bodies a `json.loads` decodes, files, WebSocket messages and standard input. Text cannot hold an operator: form fields, query parameters, headers, cookies, URL parameters, the command line, the environment, and an entry-point parameter annotated with scalars (`str`, `int`, `Optional[str]`, `Annotated[str, Query()]`, `list[str]`, …), which FastAPI validates. Limits: a Django or aiohttp request object is tainted as a whole, so `request.GET` counts like `request.body`; the keys of a dict are not followed, so a string concatenated into `$where` is not reported. The analyzer ships no NoSQL sink: model plugins declare them. |
+| `nosql-injection` | A NoSQL query, such as a MongoDB filter, given attacker input that may hold a structure: a JSON body giving `{"$ne": null}` matches any value. The input that may hold a structure is JSON, the raw bodies a `json.loads` decodes, files, WebSocket messages and standard input. Text cannot hold an operator: form fields, query parameters, headers, cookies, URL parameters, the command line, the environment, and an entry-point parameter annotated with scalars (`str`, `int`, `Optional[str]`, `Annotated[str, Query()]`, `list[str]`, …), which FastAPI validates. A Django, REST framework, aiohttp or tornado view reads a request object: its text attributes (`GET`, `POST`, `headers`, `query_params`, `query`, `match_info`, `arguments`, …) and the URL parameters after it are text, its body, files and `data` are not. Limits: a helper function the request is passed to reads it as a whole; the keys of a dict are not followed, so a string concatenated into `$where` is not reported. The analyzer ships no NoSQL sink: model plugins declare them. |
 | `path-traversal` | A file system path (`open`, `send_file`, `os.remove`, …). |
 | `ssrf` | The URL of an HTTP client request (Requests, httpx, `urllib`), by position or as `url=`; the body, headers and query parameters are not the destination. |
 | `xss` | An HTTP response body built without escaping. A Django template rendered by `render_to_string` escapes, where the analyzer can establish it (below). |
@@ -241,7 +241,7 @@ the directory of the checked file; a path outside it is printed as it is.
 ```json
 {
   "schema_version": 1,
-  "tool": {"name": "coretrace-python-analyzer", "version": "0.12.0"},
+  "tool": {"name": "coretrace-python-analyzer", "version": "0.13.0"},
   "root": "/home/me/project",
   "findings": [
     {
@@ -273,7 +273,7 @@ bytes. A single-file check lists no components.
 
 ```json
 "tool": {
-  "name": "coretrace-python-analyzer", "version": "0.12.0",
+  "name": "coretrace-python-analyzer", "version": "0.13.0",
   "components": [
     {"kind": "plugin", "name": "curated-advisories", "version": "2026.09.25", "digest": "sha256:…"},
     {"kind": "advisories", "name": "advisories.json", "digest": "sha256:…"}
@@ -436,7 +436,7 @@ project's behalf would go unseen. Without a lock file, a package is never
 `not_affected`. The document's `@id` derives from its statements, so two runs with the
 same result share it; its `author` is `Unknown Author`, as OpenVEX tools write when they
 cannot know it. Its `tooling` names the analyzer and each component with its digest
-(`coretrace-python-analyzer 0.12.0; curated-advisories 2026.09.25 (sha256:…)`), so the
+(`coretrace-python-analyzer 0.13.0; curated-advisories 2026.09.25 (sha256:…)`), so the
 `@id` changes with the advisory data a statement was decided with.
 
 ## Large projects

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "hatchling.build"
 
 [project]
 name = "coretrace-python-analyzer"
-version = "0.12.0"
+version = "0.13.0"
 description = "Static security analysis infrastructure for Python code"
 readme = "README.md"
 requires-python = ">=3.11"

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/src/coretrace_python/__init__.py

@@ -1,4 +1,4 @@
 """CoreTrace's Python static analysis frontend."""
 
-__version__ = "0.12.0"
+__version__ = "0.13.0"
 

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/src/coretrace_python/bundled/models/aiohttp/aiohttp_models.py

@@ -12,6 +12,7 @@ from coretrace_python.semantic.symbols import SymbolId
 from coretrace_python.taint import (
     EntryPoint,
     Model,
+    RequestObject,
     RouteRegistrar,
     Sink,
     Source,
@@ -22,6 +23,23 @@ from coretrace_python.taint import (
 _METHODS = ("get", "post", "put", "patch", "delete", "head", "options")
 _REQUEST_CLASSES = ("aiohttp.web.Request", "aiohttp.web_request.Request")
 _ROUTERS = ("aiohttp.web.Application.router", "aiohttp.web.UrlDispatcher")
+# ``@routes.get('/')`` on a ``RouteTableDef``.
+_ROUTE_TABLES = tuple(f"aiohttp.web.RouteTableDef.{method}" for method in (*_METHODS, "route", "view"))
+# ``app.router.add_route('GET', '/', handler)``, ``app.router.add_get('/', handler)`` and
+# ``app.add_routes([web.get('/', handler)])``, each with the argument naming the handler.
+_ROUTE_CALLS = (
+    *((f"{router}.add_route", 2) for router in _ROUTERS),
+    *((f"{router}.add_{method}", 1) for router in _ROUTERS for method in (*_METHODS, "view")),
+    ("aiohttp.web.route", 2),
+    *((f"aiohttp.web.{method}", 1) for method in (*_METHODS, "view")),
+)
+# What a request gives as text; its body (``json()``, ``text()``, ``read()``) may hold a
+# structure.
+_REQUEST_TEXT = (
+    "query", "rel_url", "url", "path", "raw_path", "path_qs", "query_string", "match_info", "headers",
+    "raw_headers", "cookies", "method", "host", "scheme", "remote", "forwarded", "content_type", "charset",
+    "if_modified_since", "post",
+)
 _REDIRECTS = ("HTTPFound", "HTTPMovedPermanently", "HTTPSeeOther", "HTTPTemporaryRedirect", "HTTPPermanentRedirect")
 _CLIENT_FUNCTIONS = (
     *(f"aiohttp.ClientSession.{method}" for method in (*_METHODS, "request")),
@@ -40,18 +58,12 @@ class AiohttpModels(ModelPlugin):
         # The whole request is attacker-controlled: its query, match info, headers,
         # cookies and body all come from the client.
         *(TypedParameter(_sym(cls), "http") for cls in _REQUEST_CLASSES),
-        # ``@routes.get('/')`` on a ``RouteTableDef``.
-        *(EntryPoint(_sym(f"aiohttp.web.RouteTableDef.{method}"), "http") for method in (*_METHODS, "route", "view")),
-        # ``app.router.add_route('GET', '/', handler)`` and ``app.router.add_get('/', handler)``.
-        *(RouteRegistrar(_sym(f"{router}.add_route"), 2, "http", keyword="handler") for router in _ROUTERS),
+        *(EntryPoint(_sym(table), "http") for table in _ROUTE_TABLES),
+        *(RouteRegistrar(_sym(call), argument, "http", keyword="handler") for call, argument in _ROUTE_CALLS),
         *(
-            RouteRegistrar(_sym(f"{router}.add_{method}"), 1, "http", keyword="handler")
-            for router in _ROUTERS
-            for method in (*_METHODS, "view")
+            RequestObject(_sym(symbol), _REQUEST_TEXT)
+            for symbol in (*_REQUEST_CLASSES, *_ROUTE_TABLES, *(call for call, _ in _ROUTE_CALLS))
         ),
-        # ``app.add_routes([web.get('/', handler)])``: the route definitions name the handler.
-        RouteRegistrar(_sym("aiohttp.web.route"), 2, "http", keyword="handler"),
-        *(RouteRegistrar(_sym(f"aiohttp.web.{method}"), 1, "http", keyword="handler") for method in (*_METHODS, "view")),
         Sink(_sym("aiohttp.web.Response"), TaintKind.HTML),
         Sink(_sym("aiohttp.web.FileResponse"), TaintKind.PATH),
         *(Sink(_sym(f"aiohttp.web.{exception}"), TaintKind.REDIRECT) for exception in _REDIRECTS),

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/src/coretrace_python/bundled/models/django/django_models.py

@@ -17,6 +17,7 @@ from coretrace_python.taint import (
     AuthorizationGuard,
     EntryPoint,
     Model,
+    RequestObject,
     RouteRegistrar,
     Sanitizer,
     Sink,
@@ -88,6 +89,23 @@ _AUTHORIZATION_DECORATORS = (
 )
 
 
+# What a request object gives as text: the query string, form fields, headers, cookies
+# and path. Its body, its files and REST framework's parsed ``data`` may hold a structure.
+_REQUEST_TEXT = (
+    "GET", "POST", "COOKIES", "META", "headers", "path", "path_info", "method", "scheme", "encoding",
+    "content_type", "content_params", "resolver_match", "get_full_path", "get_full_path_info",
+    "build_absolute_uri", "get_host", "get_port", "get_signed_cookie", "query_params",
+)
+# Every input a request object comes from: an annotated parameter, a class-based view's
+# method or ``self.request``, a decorated view, a view registered in a URL configuration.
+_REQUEST_INPUTS = (
+    *_REQUEST_CLASSES,
+    *_VIEW_BASES,
+    *(f"{base}.request" for base in _VIEW_BASES),
+    *_VIEW_DECORATORS,
+    *_ROUTE_REGISTRARS,
+)
+
 _TARGET_ONLY = ((TaintKind.REDIRECT, (0,)),)
 
 
@@ -103,6 +121,7 @@ class DjangoModels(ModelPlugin):
         # ``self.request`` in a class-based view: the attribute is inherited from the base.
         *(Source(_sym(f"{base}.request"), "http") for base in _VIEW_BASES),
         *(EntryPoint(_sym(decorator), "http") for decorator in _VIEW_DECORATORS),
+        *(RequestObject(_sym(symbol), _REQUEST_TEXT) for symbol in _REQUEST_INPUTS),
         Sink(_sym("django.db.connection.cursor.execute"), TaintKind.SQL | TaintKind.CREDENTIAL, ((TaintKind.SQL, (0,)),)),
         Sink(_sym("django.db.connection.cursor.executemany"), TaintKind.SQL | TaintKind.CREDENTIAL, ((TaintKind.SQL, (0,)),)),
         Sink(_sym("django.db.models.expressions.RawSQL"), TaintKind.SQL | TaintKind.CREDENTIAL),

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/src/coretrace_python/bundled/models/tornado/tornado_models.py

@@ -8,7 +8,16 @@ from typing import ClassVar
 
 from coretrace_python.plugins import ModelPlugin
 from coretrace_python.semantic.symbols import SymbolId
-from coretrace_python.taint import TEXT_KINDS, EntryPoint, Model, Sanitizer, Sink, Source, TaintKind
+from coretrace_python.taint import (
+    TEXT_KINDS,
+    EntryPoint,
+    Model,
+    RequestObject,
+    Sanitizer,
+    Sink,
+    Source,
+    TaintKind,
+)
 
 _HANDLERS = ("tornado.web.RequestHandler", "tornado.websocket.WebSocketHandler")
 # Text, which cannot hold a query operator; the request object holds the raw body too.
@@ -16,6 +25,11 @@ _TEXT_METHODS = (
     "get_argument", "get_arguments", "get_query_argument", "get_query_arguments",
     "get_body_argument", "get_body_arguments", "get_cookie", "path_args", "path_kwargs",
 )
+# What ``self.request`` gives as text; its body and files may hold a structure.
+_REQUEST_TEXT = (
+    "arguments", "query_arguments", "body_arguments", "query", "uri", "path", "headers", "cookies",
+    "method", "host", "host_name", "remote_ip", "protocol", "version", "full_url",
+)
 _CLIENTS = ("tornado.httpclient.AsyncHTTPClient.fetch", "tornado.httpclient.HTTPClient.fetch")
 _TARGET_ONLY = ((TaintKind.REDIRECT, (0,)),)
 
@@ -33,6 +47,7 @@ class TornadoModels(ModelPlugin):
         EntryPoint(_sym("tornado.websocket.WebSocketHandler"), "http"),
         *(Source(_sym(f"{handler}.{method}"), "http", TEXT_KINDS) for handler in _HANDLERS for method in _TEXT_METHODS),
         *(Source(_sym(f"{handler}.request"), "http") for handler in _HANDLERS),
+        *(RequestObject(_sym(f"{handler}.request"), _REQUEST_TEXT) for handler in _HANDLERS),
         *(Sink(_sym(f"{handler}.write"), TaintKind.HTML) for handler in _HANDLERS),
         *(Sink(_sym(f"{handler}.finish"), TaintKind.HTML) for handler in _HANDLERS),
         *(Sink(_sym(f"{handler}.redirect"), TaintKind.REDIRECT, _TARGET_ONLY) for handler in _HANDLERS),

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/src/coretrace_python/taint/__init__.py

@@ -17,6 +17,7 @@ from coretrace_python.taint.models import (
     ModelError,
     ModelTable,
     NamedParameter,
+    RequestObject,
     RouteRegistrar,
     SafeArgument,
     Sanitizer,
@@ -45,6 +46,7 @@ __all__ = [
     "ModelTable",
     "NamedParameter",
     "RegisteredRoutes",
+    "RequestObject",
     "RouteRegistrar",
     "Routes",
     "SafeArgument",

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/src/coretrace_python/taint/engine.py

@@ -74,6 +74,7 @@ from coretrace_python.semantic.scopes import ScopeAnalysis, ScopeId, ScopeTable
 from coretrace_python.semantic.symbols import SymbolAnalysis, SymbolId, SymbolTable
 from coretrace_python.source import SourceSpan
 from coretrace_python.taint.models import (
+    TEXT_KINDS,
     EntryPoint,
     ModelTable,
     SecurityModelAnalysis,
@@ -84,6 +85,9 @@ from coretrace_python.taint.models import (
 from coretrace_python.taint.routes import RegisteredRoutes, Routes
 from coretrace_python.taint.templates import EscapedTemplates
 
+# What only a structure carries, and text does not.
+_STRUCTURED = TaintKind.ALL & ~TEXT_KINDS
+
 
 @dataclass(frozen=True)
 class Taint:
@@ -274,19 +278,42 @@ class _TaintProblem(DataflowProblem[State]):
         taint = Taint.none()
         symbol = self.symbols.get(instruction.result) if instruction.result else None
         if symbol is not None:
-            source = self.models.source_covering(symbol)
-            if source is not None:
-                taint = Taint(source.kinds, frozenset({source}))
+            taint = self.covered(symbol)
         if isinstance(instruction, Symbol | Compare):
             return taint
         for operand in instruction.operands():
-            taint = taint.join(state.get(operand, Taint.none()))
+            operand_taint = state.get(operand, Taint.none())
+            if isinstance(instruction, GetAttr):
+                operand_taint = self.attribute(operand_taint, instruction.attribute)
+            taint = taint.join(operand_taint)
         if isinstance(instruction, BuildList | BuildTuple | BuildDict | BuildSet):
             # ``[*xs]`` and ``{**d}`` copy the contents of what they unpack.
             for unpacked in instruction.unpacked:
                 taint = taint.join(self.deep(unpacked, state))
         return taint
 
+    def covered(self, symbol: SymbolId) -> Taint:
+        """The taint of the source covering ``symbol``, if any. Below a request source, as
+        ``self.request.GET`` is, it is the taint of the request attribute read."""
+
+        source = self.models.source_covering(symbol)
+        if source is None:
+            return Taint.none()
+        taint = Taint(source.kinds, frozenset({source}))
+        if symbol == source.symbol:
+            return taint
+        read = symbol.canonical_name.removeprefix(f"{source.symbol.canonical_name}.").split(".")[0]
+        return self.attribute(taint, read)
+
+    def attribute(self, taint: Taint, name: str) -> Taint:
+        """What attribute ``name`` of a value carrying ``taint`` carries: text when every
+        source of the value is a request object listing ``name`` among its text attributes."""
+
+        requests = [self.models.request_object(source.symbol) for source in taint.sources]
+        if taint and all(request is not None and name in request.text for request in requests):
+            return taint.without(_STRUCTURED)
+        return taint
+
     def call(self, call: Call, state: dict[Key, Taint], flows: list[TaintFlow]) -> Taint:
         arguments = tuple(self.deep(a, state) for a in call.arguments)
         keywords = Taint.none()
@@ -420,11 +447,7 @@ class _TaintProblem(DataflowProblem[State]):
         if cleared:
             return everything.without(TaintKind(cleared))
         # A method on a tainted object returns tainted data (``request.args.get``).
-        everything = everything.join(state.get(call.callee, Taint.none()))
-        source = self.models.source_covering(symbol)
-        if source is not None:
-            return everything.join(Taint(source.kinds, frozenset({source})))
-        return everything
+        return everything.join(state.get(call.callee, Taint.none())).join(self.covered(symbol))
 
     def guarded(self, sink: Sink | None, arguments: Arguments) -> Sink | None:
         """The sink as this call makes it: an argument declared safe (``yaml.load`` with
@@ -777,7 +800,8 @@ def parameter_sources(
     entry point (``self`` excepted for a method), including one registered elsewhere
     (``routes``), and every parameter annotated with a typed-parameter symbol. An entry
     point's parameter annotated with scalars is text: FastAPI validates it before the
-    handler runs, and a scalar cannot hold a query operator."""
+    handler runs, and a scalar cannot hold a query operator. So are the URL parameters
+    a view receives after a request object."""
 
     owner = _owner(module, function)
     sources: dict[int, Source] = {}
@@ -786,10 +810,13 @@ def parameter_sources(
     entry = function_entry_point(function, module, models, scopes, symbols, instances, routes)
     if entry is not None:
         first = 1 if owner is not None else 0
+        request = models.request_object(entry.symbol)
         for index in range(first, len(function.parameters)):
             annotation = function.parameters[index].annotation
-            scalar = annotation is not None and _scalar(annotation, symbols, enclosing)
-            kinds = entry.kinds & ~TaintKind.NOSQL if scalar else entry.kinds
+            text = (request is not None and index > first) or (
+                annotation is not None and _scalar(annotation, symbols, enclosing)
+            )
+            kinds = entry.kinds & ~_STRUCTURED if text else entry.kinds
             sources[index] = Source(entry.symbol, entry.label, kinds)
     for index, parameter in enumerate(function.parameters):
         if parameter.annotation is None:

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/src/coretrace_python/taint/models.py

@@ -205,6 +205,22 @@ class TemplateRender:
     keyword: str = "template_name"
 
 
+@dataclass(frozen=True)
+class RequestObject:
+    """The inputs from ``symbol`` are request objects; ``symbol`` names an entry point, a
+    route registrar, a typed parameter or a source. A view receives the request first
+    (after ``self``), then URL parameters, which are text. Reading one of the request's
+    ``text`` attributes (``GET``, ``headers``) gives text, while its body and files keep the
+    kinds of the request: text carries ``TEXT_KINDS``."""
+
+    symbol: SymbolId
+    text: tuple[str, ...] = ()
+
+    def __post_init__(self) -> None:
+        # Sorted: the models' repr is part of the cache key, the same in every process.
+        object.__setattr__(self, "text", tuple(sorted(set(self.text))))
+
+
 @dataclass(frozen=True)
 class AuthorizationGuard:
     """A decorator, or a condition, that restricts who reaches the code behind it; a
@@ -228,6 +244,7 @@ Model = (
     | SafeArgument
     | TemplateRender
     | Members
+    | RequestObject
 )
 
 
@@ -246,6 +263,7 @@ class ModelTable:
     safe_arguments: tuple[SafeArgument, ...] = ()
     template_renders: tuple[TemplateRender, ...] = ()
     members: tuple[Members, ...] = ()
+    request_objects: tuple[RequestObject, ...] = ()
     _by_symbol: dict[type[Model], dict[SymbolId, Model]] = field(
         init=False, repr=False, compare=False
     )
@@ -262,6 +280,7 @@ class ModelTable:
             RouteRegistrar: {m.symbol: m for m in self.route_registrars},
             SafeArgument: {m.symbol: m for m in self.safe_arguments},
             TemplateRender: {m.symbol: m for m in self.template_renders},
+            RequestObject: {m.symbol: m for m in self.request_objects},
         }
         object.__setattr__(self, "_by_symbol", MappingProxyType(index))
 
@@ -330,6 +349,7 @@ class ModelTable:
             self.safe_arguments,
             self.template_renders,
             self.members,
+            self.request_objects,
         )
 
     def sanitizer(self, symbol: SymbolId) -> Sanitizer | None:
@@ -344,6 +364,10 @@ class ModelTable:
         found = self._by_symbol[TemplateRender].get(symbol)
         return found if isinstance(found, TemplateRender) else None
 
+    def request_object(self, symbol: SymbolId) -> RequestObject | None:
+        found = self._by_symbol[RequestObject].get(symbol)
+        return found if isinstance(found, RequestObject) else None
+
     def members_by_class(self) -> Mapping[SymbolId, Members]:
         """The ``Members`` models by class, as symbol resolution reads them."""
 
@@ -408,6 +432,7 @@ class SecurityModelRegistry:
             safe_arguments=tuple(m for m in models if isinstance(m, SafeArgument)),
             template_renders=tuple(m for m in models if isinstance(m, TemplateRender)),
             members=tuple(m for m in models if isinstance(m, Members)),
+            request_objects=tuple(m for m in models if isinstance(m, RequestObject)),
         )
 
 

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/tests/test_nosql_injection.py

@@ -11,9 +11,13 @@ text too, since FastAPI validates it before the handler runs.
 
 The engine ships no NOSQL sink; these tests declare one, ``docstore.find(filter)``.
 
-Limits: a Django or aiohttp request object is tainted as a whole, so ``request.GET``
-carries ``NOSQL`` like ``request.body``; the keys of a dict literal are not followed, so
-a string concatenated into ``$where`` is not reported.
+Django, Django REST framework, aiohttp and tornado hand views a request object, tainted as
+a whole: its text attributes (``GET``, ``headers``, ``query``, ``arguments``) are text, its
+body and files are not, and the URL parameters a view receives after it are text.
+
+Limits: a helper function the request is passed to reads it as a whole, since function
+summaries do not know where their parameters come from; the keys of a dict literal are
+not followed, so a string concatenated into ``$where`` is not reported.
 """
 
 from __future__ import annotations
@@ -176,9 +180,12 @@ def test_standard_input_is_a_payload_and_the_command_line_text(plugins: list[Pat
 
 
 DJANGO = (
-    "import json\n\nimport docstore\nfrom django.http import HttpRequest, JsonResponse\n\n"
-    "def users(request: HttpRequest):\n"
-    "    return JsonResponse(docstore.find({query}))\n"
+    "import json\n\nimport docstore\nfrom django.http import HttpRequest, JsonResponse\nfrom django.urls import path\n\n"
+    "def search(request, slug):\n"
+    "    return JsonResponse(docstore.find({query}))\n\n"
+    "def typed(request: HttpRequest):\n"
+    "    return JsonResponse(docstore.find({query}))\n\n"
+    "urlpatterns = [path('search/<slug>/', search)]\n"
 )
 
 
@@ -186,12 +193,131 @@ DJANGO = (
     "query",
     [
         "json.loads(request.body)",
-        # Limit: the request object is tainted as a whole, its text fields too.
-        '{"name": request.GET["name"]}',
+        "json.loads(request.read())",
+        "json.load(request.FILES['filter'])",
+    ],
+)
+def test_the_body_and_files_of_a_django_request_are_payloads(plugins: list[Path], query: str) -> None:
+    assert injected(plugins, DJANGO.format(query=query)) == [8, 11]
+
+
+@pytest.mark.parametrize(
+    "query",
+    [
+        "{'name': request.GET['name']}",
+        "{'name': request.POST.get('name')}",
+        "{'agent': request.headers['User-Agent'], 'meta': request.META['REMOTE_ADDR']}",
+        "{'token': request.COOKIES['token'], 'path': request.get_full_path()}",
+    ],
+)
+def test_the_text_attributes_of_a_django_request_are_text(plugins: list[Path], query: str) -> None:
+    assert injected(plugins, DJANGO.format(query=query)) == []
+
+
+def test_the_url_parameters_after_the_request_are_text(plugins: list[Path]) -> None:
+    text = DJANGO.format(query="{'tag': slug}").replace("def typed(request: HttpRequest)", "def typed(request, slug)")
+
+    assert injected(plugins, text) == []
+
+
+@pytest.mark.parametrize(
+    "query, reported",
+    [
+        ("json.loads(request.body)", True),
+        ("json.loads(self.request.body)", True),
+        ("{'name': request.POST['name']}", False),
+        ("{'name': self.request.GET['name']}", False),
+        ("{'tag': slug}", False),
     ],
 )
-def test_a_django_request_carries_nosql_as_a_whole(plugins: list[Path], query: str) -> None:
-    assert injected(plugins, DJANGO.format(query=query)) == [7]
+def test_a_class_based_view_receives_the_same_request(plugins: list[Path], query: str, reported: bool) -> None:
+    text = (
+        "import json\n\nimport docstore\nfrom django.views import View\n\n"
+        "class Search(View):\n"
+        "    def post(self, request, slug):\n"
+        f"        return docstore.find({query})\n"
+    )
+
+    assert injected(plugins, text) == ([8] if reported else [])
+
+
+@pytest.mark.parametrize("query, reported", [("request.data", True), ("{'q': request.query_params['q']}", False)])
+def test_a_rest_framework_request_parses_the_body_into_data(plugins: list[Path], query: str, reported: bool) -> None:
+    text = (
+        "import docstore\nfrom rest_framework.decorators import api_view\n\n"
+        "@api_view(['POST'])\n"
+        "def create(request):\n"
+        f"    return docstore.find({query})\n"
+    )
+
+    assert injected(plugins, text) == ([6] if reported else [])
+
+
+@pytest.mark.parametrize(
+    "query, reported",
+    [
+        ("await request.json()", True),
+        ("{'name': request.match_info['name']}", False),
+        ("{'q': request.query['q'], 'agent': request.headers['User-Agent']}", False),
+        ("{'name': (await request.post())['name']}", False),
+    ],
+)
+def test_an_aiohttp_request_is_text_but_for_its_body(plugins: list[Path], query: str, reported: bool) -> None:
+    text = (
+        "import docstore\nfrom aiohttp import web\n\n"
+        "routes = web.RouteTableDef()\n\n"
+        "@routes.post('/users/{{name}}')\n"
+        "async def users(request):\n"
+        f"    return docstore.find({query})\n\n"
+        "async def typed(request: web.Request):\n"
+        f"    return docstore.find({query})\n"
+    )
+
+    assert injected(plugins, text) == ([8, 11] if reported else [])
+
+
+@pytest.mark.parametrize(
+    "query, reported",
+    [
+        ("json.loads(self.request.body)", True),
+        ("{'args': self.request.arguments, 'uri': self.request.uri}", False),
+        ("{'name': name}", False),
+    ],
+)
+def test_a_tornado_request_is_text_but_for_its_body(plugins: list[Path], query: str, reported: bool) -> None:
+    text = (
+        "import json\n\nimport docstore\nimport tornado.web\n\n"
+        "class Users(tornado.web.RequestHandler):\n"
+        "    def post(self, name):\n"
+        f"        docstore.find({query})\n"
+    )
+
+    assert injected(plugins, text) == ([8] if reported else [])
+
+
+def test_text_attributes_still_carry_the_other_kinds(plugins: list[Path]) -> None:
+    text = (
+        "from django.http import HttpResponse\nfrom django.urls import path\n\n"
+        "def hello(request, slug):\n"
+        "    return HttpResponse(request.GET['name'] + slug)\n\n"
+        "urlpatterns = [path('hello/<slug>/', hello)]\n"
+    )
+
+    assert [f.span.start_line for f in check(plugins, text) if f.rule_id == "xss"] == [5]
+
+
+def test_a_helper_the_request_is_passed_to_reads_it_as_a_whole(plugins: list[Path]) -> None:
+    text = (
+        "import docstore\nfrom rest_framework.decorators import api_view\n\n"
+        "def lookup(request):\n"
+        "    return docstore.find({'name': request.GET['name']})\n\n"
+        "@api_view(['GET'])\n"
+        "def view(request):\n"
+        "    return lookup(request)\n"
+    )
+
+    # Limit: the summary of ``lookup`` does not know its parameter is a request.
+    assert injected(plugins, text) == [9]
 
 
 def test_parameters_of_one_entry_point_remain_one_origin(plugins: list[Path]) -> None:

{coretrace_python_analyzer-0.12.0 → coretrace_python_analyzer-0.13.0}/tests/test_release.py

@@ -30,7 +30,7 @@ def require_changelog() -> None:
 def test_the_version_is_declared_once_and_mirrored() -> None:
     project = tomllib.loads((REPO / "pyproject.toml").read_text(encoding="utf-8"))["project"]
 
-    assert project["version"] == coretrace_python.__version__ == "0.12.0"
+    assert project["version"] == coretrace_python.__version__ == "0.13.0"
 
 
 def test_the_changelog_starts_with_the_current_version() -> None: