mcp-server-semgrep/semgrep-rules/python/lang/security/dangerous-code-run.yaml

MCP Server for Semgrep Integration - static code analysis with AI

rules:
- id: dangerous-interactive-code-run
  mode: taint
  options:
    symbolic_propagation: true
  pattern-sources:
  - patterns:
    - pattern-either:
      - patterns:
        - pattern-either:
          - pattern: flask.request.form.get(...)
          - pattern: flask.request.form[...]
          - pattern: flask.request.args.get(...)
          - pattern: flask.request.args[...]
          - pattern: flask.request.values.get(...)
          - pattern: flask.request.values[...]
          - pattern: flask.request.cookies.get(...)
          - pattern: flask.request.cookies[...]
          - pattern: flask.request.stream
          - pattern: flask.request.headers.get(...)
          - pattern: flask.request.headers[...]
          - pattern: flask.request.data
          - pattern: flask.request.full_path
          - pattern: flask.request.url
          - pattern: flask.request.json
          - pattern: flask.request.get_json()
          - pattern: flask.request.view_args.get(...)
          - pattern: flask.request.view_args[...]
          - patterns:
            - pattern-inside: |
                @$APP.route(...)
                def $FUNC(..., $ROUTEVAR, ...):
                  ...
            - focus-metavariable: $ROUTEVAR
      - patterns:
        - pattern-inside: |
            def $FUNC(request, ...):
              ...
        - pattern-either:
          - pattern: request.$PROPERTY.get(...)
          - pattern: request.$PROPERTY[...]
      - patterns:
        - pattern-either:
          - pattern-inside: |
              @rest_framework.decorators.api_view(...)
              def $FUNC($REQ, ...):
                ...
          - patterns:
            - pattern-either:
              - pattern-inside: |
                  class $VIEW(..., rest_framework.views.APIView, ...):
                    ...
              - pattern-inside: |
                  class $VIEW(..., rest_framework.generics.GenericAPIView, ...):
                    ...                              
            - pattern-inside: |
                def $METHOD(self, $REQ, ...):
                  ...
            - metavariable-regex:
                metavariable: $METHOD
                regex: (get|post|put|patch|delete|head)
        - pattern-either:
          - pattern: $REQ.POST.get(...)
          - pattern: $REQ.POST[...]
          - pattern: $REQ.FILES.get(...)
          - pattern: $REQ.FILES[...]
          - pattern: $REQ.DATA.get(...)
          - pattern: $REQ.DATA[...]
          - pattern: $REQ.QUERY_PARAMS.get(...)
          - pattern: $REQ.QUERY_PARAMS[...]
          - pattern: $REQ.data.get(...)
          - pattern: $REQ.data[...]
          - pattern: $REQ.query_params.get(...)
          - pattern: $REQ.query_params[...]
          - pattern: $REQ.content_type
          - pattern: $REQ.content_type
          - pattern: $REQ.stream
          - pattern: $REQ.stream
      - patterns:
        - pattern-either:
          - pattern-inside: |
              class $SERVER(..., http.server.BaseHTTPRequestHandler, ...):
                ...
          - pattern-inside: |
              class $SERVER(..., http.server.StreamRequestHandler, ...):
                ...
          - pattern-inside: |
              class $SERVER(..., http.server.DatagramRequestHandler, ...):
                ...
        - pattern-either:
          - pattern: self.requestline
          - pattern: self.path
          - pattern: self.headers[...]
          - pattern: self.headers.get(...)
          - pattern: self.rfile
      - patterns:
        - pattern-inside: |
            @pyramid.view.view_config( ... )
            def $VIEW($REQ):
              ...
        - pattern: $REQ.$ANYTHING
        - pattern-not: $REQ.dbsession
  pattern-sinks:
  - patterns:
    - pattern-either:
      - pattern-inside: |
          $X = code.InteractiveConsole(...)
          ...
      - pattern-inside: |
          $X = code.InteractiveInterpreter(...)
          ...
    - pattern-either:
      - pattern: |
          $X.push($PAYLOAD,...)
      - pattern: |
          $X.runsource($PAYLOAD,...)
      - pattern: |
          $X.runcode(code.compile_command($PAYLOAD),...)
      - pattern: |
          $PL = code.compile_command($PAYLOAD,...)
          ...
          $X.runcode($PL,...)
    - focus-metavariable: $PAYLOAD
    - pattern-not: |
        $X.push("...",...)
    - pattern-not: |
        $X.runsource("...",...)
    - pattern-not: |
        $X.runcode(code.compile_command("..."),...)
    - pattern-not: |
        $PL = code.compile_command("...",...)
        ...
        $X.runcode($PL,...)
  message: >-
    Found user controlled data inside InteractiveConsole/InteractiveInterpreter method.
    This is dangerous if external data can reach this function call because it allows
    a malicious actor to run arbitrary Python code.
  metadata:
    cwe:
    - "CWE-95: Improper Neutralization of Directives in Dynamically Evaluated Code ('Eval Injection')"
    owasp:
    - A03:2021 - Injection
    references:
    - https://semgrep.dev/docs/cheat-sheets/python-command-injection/
    category: security
    technology:
    - python
    confidence: MEDIUM
    subcategory:
    - vuln
    likelihood: MEDIUM
    impact: HIGH
  severity: WARNING
  languages:
  - python