{% extends "base.html" %} {% import "_ui.html" as ui %} {% block title %}Runs{% endblock %} {% block crumbs %}Runs{% endblock %} {% block topbar_actions %}{% endblock %} {% block content %} {%- set n_rules = runs|sum(attribute="rules") -%} {%- set n_approved = runs|sum(attribute="approved") -%} {% if deleted %}
{{ ui.icon("check-circle") }}
Run deleted
{{ deleted }}
{% endif %}

Detection runs

Turn a threat-intel or incident report into ATT&CK-mapped Sigma rules, SIEM queries for every backend, and safe simulation guides, then review, test and approve them here.

{{ ui.stat("Reports analysed", runs|length, "file") }} {{ ui.stat("Rules generated", n_rules, "target", hint=(runs|sum(attribute="behaviours")) ~ " behaviours extracted") }} {{ ui.stat("Approved rules", n_approved, "check-circle", "ok", sub=("of " ~ n_rules) if n_rules else "", hint="Four-eyes: approver ≠ author") }} {% if ai.enabled %} {{ ui.stat("AI usage, all time", (ai.in_total + ai.out_total)|tok, "ai", sub="tokens", hint="est. $" ~ ("%.2f"|format(ai.cost_total or 0)) ~ " · " ~ ai.provider) }} {% else %} {{ ui.stat("AI", "Off", "shield", "ok", hint="Deterministic mode: nothing is sent out") }} {% endif %}

{{ ui.icon("plus") }}New analysis

Upload a report or point to a public URL. Behaviours are extracted first; you pick which become rules.

{% if ai.blocked %}
{{ ui.icon("pause") }}
The AI is paused
New analyses need the AI and will not start until the pause clears. See why
{% endif %}
Report source
Document type
{{ ui.icon("chevron-right", "sm chev") }}Advanced options
{{ ui.icon("shield") }}
The report is condensed locally before anything is sent to the model. Usernames, hosts, internal IPs, emails, SIDs, secrets and configured customer names are redacted first, and the TLP policy is enforced. Simulations are generated, linted and documented, never executed by this tool.
You can follow progress live; nothing is lost if you close the tab.

{{ ui.icon("clock") }}Recent jobs

{% if jobs %}
    {% for j in jobs %}
  • {{ j.label }} {{ ui.short_date(j.created_at) }} · {{ j.actor }} {{ ui.status(j.status, dot=False) }}
  • {% endfor %}
{% else %}{{ ui.empty("No jobs yet", "Analyses and AI fixes run in the background and show up here.", "clock") }}{% endif %}

{{ ui.icon("database") }}SIEM coverage {{ siems|length }}

Every rule is converted for each of these. Set them in ti2dac.yaml.

{% if siems %}
{% for label, lang, ai_made in siems %} {{ lang }}{{ label.split(' (')[0] }}{% if '(' in label %} {{ label.split(' (', 1)[1].rstrip(')') }}{% endif %} {% endfor %}

Deterministic pySigma conversion costs no tokens.{% if siems|selectattr(2)|list %} Amber targets are AI-translated and unverified.{% endif %}

{% else %}{{ ui.empty("No SIEM targets", "Add targets in ti2dac.yaml.", "database") }}{% endif %}

{{ ui.icon("map") }}How a run works

  1. Condense & redact locallySections, IOCs and sensitive values are stripped before the model sees anything.
  2. Extract behavioursTTPs mapped to ATT&CK, with evidence quotes checked against the source.
  3. Generate SigmaValidated with pySigma, auto-repaired, compared with SigmaHQ for duplicates.
  4. Convert & self-testQueries for every configured SIEM, plus a replay against expected events.
  5. Review & feed backApprove with four-eyes, record lab tests, and turn corrections into lessons.

{{ ui.icon("runs") }}All runs {{ runs|length }}

Most recent first. Open a run to review its behaviours and rules.

{% if runs %}{% endif %}
{% if runs %}
{% for r in runs %}{% endfor %}
ReportRulesAI reviewCreatedAI callsTokens in / outEst. costActions
{{ r.title }}
{% if r.kind == "incident" %}incident{% endif %} {% if r.tlp %}{{ r.tlp }}{% endif %} {% set src = r.source or "" %}{{ src.split('/')[2] if src.startswith('http') and src.count('/') >= 2 else src }}
{% if r.rules %}{{ r.rules }} / {{ r.behaviours }} behaviours {% if r.approved %}
{{ r.approved }} approved
{% endif %} {% else %}{{ r.behaviours }} behaviours, no rules{% endif %}
{{ ui.review_badge(r.review, r.run_id in reviewing) }} {{ ui.short_date(r.created_at) }} {{ r.usage.calls or 0 }}{% if r.usage.cached_calls %} (+{{ r.usage.cached_calls }} cached){% endif %} {{ r.usage.input_tokens|tok }} / {{ r.usage.output_tokens|tok }} {% if r.usage.estimated_cost_usd %}${{ "%.4f"|format(r.usage.estimated_cost_usd) }}{% else %}-{% endif %}
{{ ui.icon("download") }} {{ ui.icon("arrow-right", "sm") }} {% if role == "admin" or user == r.created_by %}
{% endif %}
{% else %} {{ ui.empty("No runs yet", "Analyse your first report above. It takes a minute or two.", "runs") }} {% endif %}
{% endblock %}