feat: [cape_sandbox] add sample submission and report enrichment

documentation/expansion/cape_sandbox.md

@@ -0,0 +1,11 @@
+# CAPE Sandbox Enrichment
+
+Submit supplied samples to CAPEv2 or retrieve a completed JSON report by numeric task ID. Set api_url to the API v2 base, for example https://cape.example/apiv2. api_key is optional when the server permits unauthenticated requests. submit_samples defaults to false. max_size_mb defaults to 25 (1–100). Submission returns task IDs; use a text attribute containing a task ID for a later report lookup. Completed reports yield file hashes, network pivots and extracted configurations.
+
+Input: text, attachment, malware-sample.
+
+Configuration: api_url, api_key, submit_samples, max_size_mb, max_results. Required: api_url.
+
+max_results defaults to 25 and accepts 1–100. New attributes use `to_ids=false`; the source UUID, detection flag and metadata are preserved. Objects reference the source attribute. Requests verify TLS, reject redirects and use connection/read timeouts of 5/30 seconds. API errors do not echo credentials. No automatic submission retries or background polling occur.
+
+Reference: https://capev2.readthedocs.io/en/latest/usage/api.html

misp_modules/modules/expansion/cape_sandbox.py

@@ -0,0 +1,270 @@
+import base64
+import hashlib
+import io
+import ipaddress
+import json
+import re
+import zipfile
+from urllib.parse import urlsplit
+from uuid import UUID
+import requests
+from pymisp import MISPEvent, MISPObject
+from pymisp.exceptions import PyMISPError
+
+
+class ModuleError(ValueError):
+    """A diagnostic safe to show to the analyst."""
+
+
+def _integer(config, key, default, maximum=100):
+    value = config.get(key)
+    if value is None or value == "":
+        return default
+    if isinstance(value, bool) or not str(value).isdigit() or (not 1 <= int(value) <= maximum):
+        raise ModuleError(f"{key} must be an integer between 1 and {maximum}.")
+    return int(value)
+
+
+def _enabled(config, key, default=False):
+    value = config.get(key)
+    if value is None or value == "":
+        return default
+    if isinstance(value, bool):
+        return value
+    if str(value).lower() not in ("true", "false", "1", "0"):
+        raise ModuleError(f"{key} must be true or false.")
+    return str(value).lower() in ("true", "1")
+
+
+def _url(value):
+    if not isinstance(value, str) or any(ord(c) <= 32 or ord(c) == 127 for c in value):
+        raise ModuleError("Invalid HTTPS URL.")
+    parsed = urlsplit(value)
+    if parsed.scheme != "https" or not parsed.hostname or parsed.username or parsed.password or parsed.fragment:
+        raise ModuleError("An HTTPS URL without credentials or fragment is required.")
+    parsed.port
+    return value.rstrip("/")
+
+
+def _origin(config):
+    value = _url(config.get("api_url", ""))
+    if urlsplit(value).query:
+        raise ModuleError("api_url cannot contain a query string.")
+    return value
+
+
+def _normalise(kind, value):
+    if not isinstance(value, str) or not value.strip() or any(ord(c) < 32 for c in value):
+        raise ModuleError("A nonempty string value is required.")
+    value = value.strip()
+    if kind in ("md5", "sha1", "sha256", "x509-fingerprint-sha256"):
+        length = {"md5": 32, "sha1": 40}.get(kind, 64)
+        if not re.fullmatch("[a-fA-F0-9]{%d}" % length, value):
+            raise ModuleError("Invalid hash.")
+        return value.lower()
+    if kind in ("ip-src", "ip-dst"):
+        return str(ipaddress.ip_address(value))
+    if kind in ("domain", "hostname"):
+        value = value.rstrip(".").encode("idna").decode("ascii").lower()
+        if len(value) > 253 or not all(
+            re.fullmatch("[a-z0-9](?:[a-z0-9-]{0,61}[a-z0-9])?", p) for p in value.split(".")
+        ):
+            raise ModuleError("Invalid DNS name.")
+    if kind == "vulnerability":
+        value = value.upper()
+        if not re.fullmatch("CVE-[0-9]{4}-[0-9]{4,}", value):
+            raise ModuleError("A CVE identifier is required.")
+    return value
+
+
+def _call(method, url, **kwargs):
+    try:
+        response = requests.request(method, url, timeout=(5, 30), verify=True, allow_redirects=False, **kwargs)
+    except requests.RequestException:
+        raise ModuleError(
+            "The service request failed or timed out; check the service before repeating a submission."
+        ) from None
+    if response.status_code in (204, 404) and method == "GET":
+        return None
+    if response.status_code not in (200, 201, 202):
+        raise ModuleError(f"The service returned HTTP {response.status_code}.")
+    try:
+        return response.json()
+    except ValueError:
+        raise ModuleError("The service returned invalid JSON.") from None
+
+
+def _object(event, attribute, name):
+    obj = MISPObject(name)
+    obj.add_reference(attribute["uuid"], "describes")
+    event.add_object(obj)
+    return obj
+
+
+def _annotation(event, attribute, title, data):
+    obj = _object(event, attribute, "annotation")
+    obj.comment = title
+    obj.add_attribute("text", value=json.dumps(data, sort_keys=True), to_ids=False, disable_correlation=True)
+    return obj
+
+
+def _pivot(event, attribute, kind, value):
+    value = _normalise(kind, value)
+    if kind == attribute["type"] and value == _normalise(attribute["type"], attribute["value"]):
+        return
+    obj = _object(
+        event, attribute, "file" if kind in ("md5", "sha1", "sha256") else ("url" if kind == "url" else "domain-ip")
+    )
+    relation = "ip" if kind in ("ip-src", "ip-dst") else kind
+    obj.add_attribute(relation, value=value, to_ids=False)
+
+
+def _sandbox(event, attribute, payload, link=None):
+    obj = _object(event, attribute, "sandbox-report")
+    obj.add_attribute("sandbox-type", value="on-premise", to_ids=False)
+    obj.add_attribute("on-premise-sandbox", value="cape", to_ids=False)
+    obj.add_attribute("results", value=json.dumps(payload, sort_keys=True), to_ids=False, disable_correlation=True)
+    if link:
+        obj.add_attribute("permalink", value=link, to_ids=False)
+    return obj
+
+
+def _binary(attribute, config):
+    limit = _integer(config, "max_size_mb", 25, 100) * 1024 * 1024
+    encoded = attribute.get("data")
+    if not isinstance(encoded, str) or len(encoded) > 2 * limit:
+        raise ModuleError("Missing or oversized base64 sample data.")
+    content = base64.b64decode(encoded, validate=True)
+    filename = attribute["value"]
+    if attribute["type"] == "malware-sample":
+        filename, separator, digest = filename.rpartition("|")
+        if not separator or not re.fullmatch(r"[a-fA-F0-9]{32}", digest):
+            raise ModuleError("A malware-sample value must contain filename|md5.")
+        with zipfile.ZipFile(io.BytesIO(content)) as archive:
+            entries = [
+                e for e in archive.infolist() if not e.is_dir() and e.filename != digest.lower() + ".filename.txt"
+            ]
+            if len(entries) != 1 or entries[0].file_size > limit:
+                raise ModuleError("A malware-sample archive must contain one payload within max_size_mb.")
+            content = archive.read(entries[0], pwd=b"infected")
+        if hashlib.md5(content).hexdigest() != digest.lower():
+            raise ModuleError("The unpacked sample does not match the declared MD5.")
+    filename = filename.replace("\\", "/").rsplit("/", 1)[-1]
+    if not filename or not content or len(content) > limit:
+        raise ModuleError("The sample is empty or exceeds max_size_mb.")
+    return filename, content, hashlib.sha256(content).hexdigest()
+
+
+def handler(q=False):
+    if q is False:
+        return False
+    try:
+        request = json.loads(q)
+        if not isinstance(request, dict) or not isinstance(request.get("attribute"), dict):
+            raise ModuleError("An attribute containing type, value and uuid is required.")
+        attribute = request["attribute"]
+        UUID(attribute["uuid"])
+        if attribute["type"] not in mispattributes["input"]:
+            raise ModuleError("Unsupported attribute type.")
+        value = _normalise(attribute["type"], attribute["value"])
+        config = request.get("config") or {}
+        if not isinstance(config, dict):
+            raise ModuleError("The config must be an object.")
+        for key in REQUIRED:
+            if not isinstance(config.get(key), str) or not config[key].strip():
+                raise ModuleError(f"Missing {key} configuration.")
+        limit = _integer(config, "max_results", 25)
+        event = MISPEvent()
+        event.add_attribute(**attribute)
+        _expand(event, attribute, value, config, limit)
+        result = json.loads(event.to_json())
+        return {"results": {key: result.get(key, []) for key in ("Attribute", "Object")}}
+    except ModuleError as exc:
+        return {"error": str(exc)}
+    except (
+        ValueError,
+        TypeError,
+        KeyError,
+        AttributeError,
+        IndexError,
+        OSError,
+        RuntimeError,
+        zipfile.BadZipFile,
+        PyMISPError,
+    ):
+        return {"error": "Invalid request or service response."}
+
+
+def introspection():
+    return mispattributes
+
+
+def version():
+    return dict(moduleinfo, config=moduleconfig)
+
+
+moduleinfo = {
+    "version": "1.0",
+    "author": "MISP Project",
+    "name": "CAPE Sandbox Enrichment",
+    "description": (
+        "Submit supplied samples to CAPEv2 or retrieve a completed JSON report by numeric task ID. Set api_url to the"
+        " API v2 base, for example https://cape.example/apiv2. api_key is optional when the server permits"
+        " unauthenticated requests. submit_samples defaults to false. max_size_mb defaults to 25 (1–100). Submission"
+        " returns task IDs; use a text attribute containing a task ID for a later report lookup. Completed reports"
+        " yield file hashes, network pivots and extracted configurations."
+    ),
+    "module-type": ["expansion"],
+    "data-source": "external",
+    "logo": "",
+    "requirements": ["requests", "pymisp"],
+    "features": (
+        "Submit supplied samples to CAPEv2 or retrieve a completed JSON report by numeric task ID. Set api_url to the"
+        " API v2 base, for example https://cape.example/apiv2. api_key is optional when the server permits"
+        " unauthenticated requests. submit_samples defaults to false. max_size_mb defaults to 25 (1–100). Submission"
+        " returns task IDs; use a text attribute containing a task ID for a later report lookup. Completed reports"
+        " yield file hashes, network pivots and extracted configurations."
+    ),
+    "references": ["https://capev2.readthedocs.io/en/latest/usage/api.html"],
+    "input": "text, attachment, malware-sample",
+    "output": "MISP objects linked to the original attribute.",
+}
+mispattributes = {"input": ["text", "attachment", "malware-sample"], "format": "misp_standard"}
+moduleconfig = ["api_url", "api_key", "submit_samples", "max_size_mb", "max_results"]
+REQUIRED = ["api_url"]
+
+
+def _expand(event, attribute, value, config, limit):
+    base = _origin(config)
+    headers = {"Authorization": "Token " + config["api_key"]} if config.get("api_key") else {}
+    if attribute["type"] in ("attachment", "malware-sample"):
+        if not _enabled(config, "submit_samples"):
+            raise ModuleError("Set submit_samples=true to submit a sample to CAPE.")
+        filename, content, digest = _binary(attribute, config)
+        data = _call("POST", base + "/tasks/create/file/", headers=headers, files={"file": (filename, content)})
+        if not isinstance(data, dict) or data.get("error") or (not data.get("data", {}).get("task_ids")):
+            raise ModuleError("CAPE did not acknowledge a submission; check the service before retrying.")
+        _annotation(
+            event, attribute, "CAPE submission receipt", {"task_ids": data["data"]["task_ids"], "sha256": digest}
+        )
+        return
+    if not re.fullmatch("[0-9]+", value):
+        raise ModuleError("text input must contain a numeric CAPE task ID.")
+    data = _call("GET", base + "/tasks/get/report/" + value + "/json/", headers=headers)
+    if data is None:
+        return
+    if not isinstance(data, dict) or data.get("error"):
+        raise ModuleError("CAPE report is unavailable or the task is not complete.")
+    _sandbox(event, attribute, data)
+    file_data = data.get("target", {}).get("file", {})
+    for kind in ("md5", "sha1", "sha256"):
+        if file_data.get(kind):
+            _pivot(event, attribute, kind, file_data[kind])
+    network = data.get("network", {})
+    for host in network.get("hosts", [])[:limit]:
+        _pivot(event, attribute, "ip-dst", host["ip"] if isinstance(host, dict) else host)
+    for record in network.get("dns", [])[:limit]:
+        if record.get("request"):
+            _pivot(event, attribute, "domain", record["request"])
+    if data.get("CAPE", {}).get("configs"):
+        _annotation(event, attribute, "CAPE extracted configurations", data["CAPE"]["configs"][:limit])