{% extends "base.html" %} {% block title %}Data sources — {{ site_name }}{% endblock %} {% block content %}

Data sources

Every source is queried in parallel and weighted by how strongly it asserts a verdict. Sources marked context add enrichment only — they never change the verdict.

{% for s in sources %} {% endfor %}
SourceSupportsCategoryAuthStatus
{% if s.homepage %}{{ s.display }} {% else %}{{ s.display }}{% endif %}
{{ s.name }}
{{ s.kinds|join(', ') }} {{ s.category }} {{ 'none' if s.keyless else (s.env_key or 'none') }} {# green is reserved for verdicts and for real health; "enabled" is not a finding, so an active source is a plain neutral tag. #} {% if s.state == 'active' %}active {% elif s.state == 'disabled' %}off {% else %}needs key{% endif %} {% if s.state == 'disabled' and s.opt_in %}
set {{ s.opt_in }}=1
{% endif %} {% if s.quota %}
{% if s.quota.limit is not none %}{{ s.quota.used }}/{{ s.quota.limit }} today {% else %}{{ s.quota.used }} today{% endif %} {% if s.quota.shared > 1 %} · shared by {{ s.quota.shared }} sources{% endif %} {% if s.quota.exhausted %} · spent{% endif %}
{% endif %}

Enabling keyed sources

Set the environment variable and restart the service. Keys are never sent to the browser or logged.

# /etc/threattrove/threattrove.env
ABUSECH_AUTH_KEY=...      # auth.abuse.ch — MalwareBazaar, ThreatFox, URLhaus
VT_API_KEY=...            # virustotal.com — files, URLs, domains, IPs
MD_API_KEY=...            # metadefender.com — files, URLs, domains, IPs
OTX_API_KEY=...           # otx.alienvault.com (optional, raises limits)
URLSCAN_API_KEY=...       # urlscan.io (optional, raises limits)
HA_API_KEY=...            # hybrid-analysis.com
MALSHARE_API_KEY=...      # malshare.com
TRIAGE_API_KEY=...        # tria.ge
GSB_API_KEY=...           # Google Safe Browsing v4 (non-commercial use)
PHISHTANK_APPKEY=...      # phishtank.com

The free sources (CIRCL hashlookup, AlienVault OTX, urlscan.io search, DNS, RDAP and Certificate Transparency) answer without any key, so a fresh install is useful immediately. Reverse IP lookups are free but rate limited, so they stay off until you set TT_ENABLE_REVERSE_IP=1.

Shared daily allowance

The three VirusTotal sources spend one key's allowance between them: a lookup of a hash, a URL and a domain is three requests. The app keeps a per-day counter and stops asking once the budget is gone, so a busy day ends with “budget spent” answers instead of rate-limit errors. Raise or lower it with TT_VT_DAILY_BUDGET (default 400; the free key allows 500).

VirusTotal's domain and IP lookup is off by default (TT_VT_DOMAINS=1 turns it on). A stray engine flag on a shared IP address or a popular domain says almost nothing, so this endpoint mostly buys noise, and the allowance is better spent on files and URLs.

URL blocklists

Google Safe Browsing and PhishTank are blocklists, not clearances. A URL that neither lists is reported as “no match” with no verdict, the same as a source with no key: not being on a list is not a statement that a URL is safe. Safe Browsing's API is licensed for non-commercial use only.

PhishTank currently sits behind a Cloudflare challenge that refuses non-browser clients, so in practice it answers “upstream blocks API clients (Cloudflare)” and makes no claim either way. When it does answer, only a record that is both valid and verified counts as phishing.

Second opinion

The MetaDefender sources spend their own key and their own allowance: one counter for the file, URL and domain/IP endpoints together, default 800 a day (TT_MD_DAILY_BUDGET, the community key allows 1000) and one call every 7.2 seconds (TT_MD_MIN_INTERVAL, the key allows 500 an hour).

Two engine panels are still one kind of evidence. VirusTotal and MetaDefender agree or they do not, but neither can convict a file on its own: both count as the same engine panel authority, and a malicious verdict still needs a second, independent authority — a sample feed, a sandbox, or your own blocklist.

{% endblock %}