From a4dcd00cc7f05848dcca8a018a4d17960afbe037 Mon Sep 17 00:00:00 2001 From: Guillaume BOURGEOIS Date: Sun, 12 Jul 2026 22:56:35 +0200 Subject: [PATCH] feat(feed): add URLhaus feed connector (abuse.ch malware URL feed) --- integrations/feed-urlhaus/manifest.yaml | 45 +++++++++++ .../feed-urlhaus/scripts/fetch_indicators.py | 80 +++++++++++++++++++ .../feed-urlhaus/scripts/test_connection.py | 54 +++++++++++++ 3 files changed, 179 insertions(+) create mode 100644 integrations/feed-urlhaus/manifest.yaml create mode 100644 integrations/feed-urlhaus/scripts/fetch_indicators.py create mode 100644 integrations/feed-urlhaus/scripts/test_connection.py diff --git a/integrations/feed-urlhaus/manifest.yaml b/integrations/feed-urlhaus/manifest.yaml new file mode 100644 index 0000000..d988dea --- /dev/null +++ b/integrations/feed-urlhaus/manifest.yaml @@ -0,0 +1,45 @@ +id: feed_urlhaus +name: URLhaus Feed +version: 1.0.0 +description: "abuse.ch URLhaus feed connector — pull the bulk list of malware-distribution URLs (recent additions or currently-online) and emit normalized IOCs (URL + type, with threat/tags/status context) for import into the Threat Indicator Manager. abuse.ch now requires a free Auth-Key for downloads; stdlib-only, no extra Python dependencies." +changelog: "1.0.0 — Initial release: fetch the URLhaus recent/online CSV feed." +category: feed + +# URLhaus is an abuse.ch feed. As of 2024 abuse.ch requires a free Auth-Key +# (sent as the Auth-Key header) for feed downloads. +config_schema: + properties: + api_key: + type: string + description: "abuse.ch Auth-Key (free — from your abuse.ch account profile)" + x-soar-sensitive: true + list: + type: string + description: "Which feed: recent (last additions) or online (currently active). Default recent" + default: "recent" + insecure: + type: boolean + description: "Trust any TLS certificate (not secure)" + default: false + required: [] + +commands: + - id: fetch_indicators + name: feed-urlhaus-fetch-indicators + description: "Fetch the URLhaus CSV feed and return normalized indicators." + risk: read + inputs_schema: + properties: + list: { type: string, description: "recent | online (overrides the config default)" } + max_indicators: { type: number, description: "Max indicators to return (0 = no limit, default 0)" } + required: [] + outputs_schema: { properties: {} } + + - id: test_connection + name: feed-urlhaus-test-connection + description: "Verify the URLhaus feed is reachable (used by the Test button)." + risk: read + inputs_schema: + properties: {} + required: [] + outputs_schema: { properties: {} } diff --git a/integrations/feed-urlhaus/scripts/fetch_indicators.py b/integrations/feed-urlhaus/scripts/fetch_indicators.py new file mode 100644 index 0000000..6c62a49 --- /dev/null +++ b/integrations/feed-urlhaus/scripts/fetch_indicators.py @@ -0,0 +1,80 @@ +import json, os, sys, ssl, io, csv, urllib.request, urllib.error + +BASE = "https://urlhaus.abuse.ch/downloads/" + + +def _cfg(): + return json.loads(os.environ.get("INTEGRATION_SECRETS", "{}")) + + +def _inputs(): + return json.loads(os.environ.get("INTEGRATION_INPUTS", "{}")) + + +def _ctx(cfg): + if cfg.get("insecure"): + c = ssl.create_default_context() + c.check_hostname = False + c.verify_mode = ssl.CERT_NONE + return c + return None + + +def _headers(cfg): + h = {"User-Agent": "Riposte-SOAR"} + k = cfg.get("api_key") + if k: + h["Auth-Key"] = str(k) + return h + + +def _get(url, cfg): + req = urllib.request.Request(url, headers=_headers(cfg)) + with urllib.request.urlopen(req, timeout=120, context=_ctx(cfg)) as r: + return r.read() + + +def _run(fn): + try: + print(json.dumps(fn(_cfg(), _inputs()))) + except urllib.error.HTTPError as e: + print(json.dumps({"error": "HTTP " + str(e.code), "detail": e.read().decode("utf-8", "replace")})) + sys.exit(1) + except Exception as e: + print(json.dumps({"error": str(e)})) + sys.exit(1) + + +def main(cfg, inputs): + which = str(inputs.get("list") or cfg.get("list") or "recent").lower() + fname = "csv_online" if which == "online" else "csv_recent" + raw = _get(BASE + fname + "/", cfg).decode("utf-8", "replace") + + maxn = int(inputs.get("max_indicators") or 0) + out = [] + for row in csv.reader(io.StringIO(raw)): + if not row or row[0].lstrip().startswith("#"): + continue + if row[0].strip().strip('"') == "id": # header row + continue + # columns: id,dateadded,url,url_status,last_online,threat,tags,urlhaus_link,reporter + if len(row) < 3: + continue + url = row[2].strip() + if not url: + continue + item = {"value": url, "type": "url"} + if len(row) > 3: + item["status"] = row[3].strip() + if len(row) > 5: + item["threat"] = row[5].strip() + if len(row) > 6: + item["tags"] = row[6].strip() + out.append(item) + if maxn and len(out) >= maxn: + break + + return {"source": "urlhaus:" + fname, "count": len(out), "indicators": out} + + +_run(main) diff --git a/integrations/feed-urlhaus/scripts/test_connection.py b/integrations/feed-urlhaus/scripts/test_connection.py new file mode 100644 index 0000000..66b4f82 --- /dev/null +++ b/integrations/feed-urlhaus/scripts/test_connection.py @@ -0,0 +1,54 @@ +import json, os, sys, ssl, urllib.request, urllib.error + +BASE = "https://urlhaus.abuse.ch/downloads/" + + +def _cfg(): + return json.loads(os.environ.get("INTEGRATION_SECRETS", "{}")) + + +def _inputs(): + return json.loads(os.environ.get("INTEGRATION_INPUTS", "{}")) + + +def _ctx(cfg): + if cfg.get("insecure"): + c = ssl.create_default_context() + c.check_hostname = False + c.verify_mode = ssl.CERT_NONE + return c + return None + + +def _headers(cfg): + h = {"User-Agent": "Riposte-SOAR"} + k = cfg.get("api_key") + if k: + h["Auth-Key"] = str(k) + return h + + +def _get(url, cfg): + req = urllib.request.Request(url, headers=_headers(cfg)) + with urllib.request.urlopen(req, timeout=60, context=_ctx(cfg)) as r: + return r.read() + + +def _run(fn): + try: + print(json.dumps(fn(_cfg(), _inputs()))) + except urllib.error.HTTPError as e: + print(json.dumps({"error": "HTTP " + str(e.code), "detail": e.read().decode("utf-8", "replace")})) + sys.exit(1) + except Exception as e: + print(json.dumps({"error": str(e)})) + sys.exit(1) + + +def main(cfg, inputs): + raw = _get(BASE + "csv_recent/", cfg).decode("utf-8", "replace") + n = sum(1 for ln in raw.splitlines() if ln and not ln.lstrip().startswith("#")) + return {"ok": True, "sample_lines": max(0, n - 1)} + + +_run(main)