Mythal
positioning · internal
★ INTERNAL · NOT FOR DISTRIBUTION 01 / 12

Mythal — the response, in their language.

An internal positioning document. Decoding a real inquiry from an Infrastructure and Cybersecurity leadership team, mapping their four-step workflow onto Mythal's twelve-agent pipeline, identifying the gaps in their literal ask, and preparing how we respond when other vendors are surveying the same brief. Read this before drafting the reply.

Mythal · the product name we're moving to. Codebase / repo still reads Mythal until renamed. F T ?

02 / 12THE LITERAL ASK

What they actually said.

Verbatim. Six phrases do the work. Highlighted below.

— Infrastructure & Cybersecurity leadership · vendor outreach
"I've been reaching out to all our vendors with what they offer from creating a Middleware AI Agent. This is the basic workflow: In the wake of the release of the Mythos, we are currently examining workflows for maintaining the uptime of systems, servers, and platforms. Essentially, this involves receiving a trigger for a critical CVE or similar incident, assessing the threat level, downloading the appropriate patch, and subsequently updating the system."
Six phrases to remember verbatim when we reply. They've internalized the Mythos thesis (good — our naming becomes literal). They're calling it "Middleware AI Agent" (good — positions us as glue, not another scanner). They've named four workflow steps explicitly (their mental model is sequential). They haven't yet named OT, compliance, rollback, or approvers (the gaps we land on later).
03 / 12DECODING THE ASK

What each phrase actually means in our domain.

Their language is precise but compact. Translate every phrase into the operational reality it implies.

"TRIGGER FOR A CRITICAL CVE"

Scanner ingest + intel uplift

Source: scanner finding, KEV uplift, vendor PSIRT, pre-disclosure feed. Frequency: 1,200+ CVEs/month at Class I scale. Need: a normalized, deduped canonical event with priority context.

"ASSESSING THE THREAT LEVEL"

Multi-signal prioritization

Not a single check. Combine CVSS, EPSS, KEV, exploit-in-wild, asset criticality, business impact, blast radius, change risk — output a prioritized work item with risk envelope.

"DOWNLOADING THE APPROPRIATE PATCH"

Locate · validate · score

Find vendor fix (or community patch / workaround / compensating control). Score for source authority and reliability. Sometimes the appropriate response is not a patch.

"UPDATING THE SYSTEM"

Apply · verify · close · prove

Drive the right tool for the asset class. Rescan. Run health probes. Roll back if it failed. Record evidence tagged to compliance controls.

They said "systems, servers, and platforms" — IT language. The first thing we'll surface that they didn't name: what happens when "the system" is a Siemens RTU at substation 14?

04 / 12★ THE KEY SLIDE · MAPPING

Their four steps → Mythal's twelve agents.

This is the slide that wins the conversation. Show them their own workflow, in their own order, executed by named specialists with audit trails.

STEP 1 · THEIR WORDS"Receiving a trigger for a critical CVE or similar incident"
MYTHAL · AGENTSScanner Liaison normalizes from Qualys / Tenable / Wiz / Defender / Claroty / Nozomi / Dragos. Threat Intel Aggregator enriches with KEV, EPSS, vendor PSIRT, exploit signals.
+ BEYOND THEIR ASKDedup across multiple scanners on (asset, cve). Pre-disclosure feed support (act before the public CVE drops). Prompt-injection defense on advisory bodies.
STEP 2 · THEIR WORDS"Assessing the threat level"
MYTHAL · AGENTSImpact Analyst joins CMDB + dependency graph. Change Risk scores historical failure rate for the asset class. OT Safety Officer reviews if OT/CCS.
+ BEYOND THEIR ASKBusiness-impact profile (downstream count, data sensitivity, network exposure). OT veto path with compensating-control proposal. Blast-radius scoring against asset dependency graph.
STEP 3 · THEIR WORDS"Downloading the appropriate patch"
MYTHAL · AGENTSPatch Hunter locates vendor fix, computes PatchReliabilityScore 0–1. Remediation Planner builds executable plan with rollback. Policy Gate routes by 7 rules.
+ BEYOND THEIR ASK"The appropriate patch" sometimes isn't a patch — it's an ACL tightening, an IPS virtual patch, or a monitored isolation. The Planner picks the correct response by asset class, not the obvious one.
STEP 4 · THEIR WORDS"Updating the system"
MYTHAL · AGENTSExecutor drives the right tool (Ansible / SCCM / Tanium / Panorama / OT-native). Verifier rescans, runs health probe, exploit retest. Compliance Reporter emits evidence units.
+ BEYOND THEIR ASKAutomatic rollback if Verifier rejects. HMAC-signed reasoning trace = audit log. Evidence units tagged to TSA / NIST / IEC controls — the same artifact engineer and auditor both read.
05 / 12WHAT'S NOT IN THE LITERAL ASK

Six things they will need within thirty days of saying yes.

Their four steps are correct but incomplete. Every production deployment surfaces these within a sprint. We bring them up unprompted — that's the difference between "vendor" and "partner."

GAP 1 · OT SAFETY

When "the system" is a substation RTU

Their phrasing ("systems, servers, platforms") is IT-shaped. The first OT incident — a Siemens or Wabtec or Rockwell advisory — proves the workflow needs a veto path. Mythal ships with it by default.

GAP 2 · ROLLBACK

What happens when the patch breaks the asset

"Updating the system" assumes success. In our IT data, ~3% of patches fail; on OT, ~18%. Without an automated rollback path the agent becomes a liability. Mythal denies any plan without a validated rollback (rule SG-POL-006).

GAP 3 · APPROVALS

Who signs · when

"Middleware AI Agent" implies autonomy. Real CISO approval requires human gates at named decisions. Mythal's policy gate routes to auto-apply / single-approval / dual-approval / deny — visible, versioned, enforced.

GAP 4 · AUDIT EVIDENCE

TSA, NIST, IEC — generated, not assembled

The compliance question lands within a month. Their auditor needs control-by-control evidence. Mythal emits it as a byproduct of every closed plan — not a separate workstream.

GAP 5 · INVENTORY LAYER

Before any CVE drops

Their workflow is reactive (CVE → fix). The twelfth agent, Inventory Insights, is proactive — surfaces EOL software, version sprawl, shadow IT, CCS-without-owner. The risk that no scanner has yet reported.

GAP 6 · INSURANCE / ROI

The CFO question

"Maintaining uptime" is the technical pitch. The board pitch is cyber-insurance attestation, reduced FTE on patch SLA, and breach-cost avoidance. We arrive with the framing on day one.

06 / 12THEIR FRAMING IS A GIFT

"Middleware AI Agent" is the right language — and what it understates.

Adopt their framing in our reply. Then extend it. Three dimensions where the phrase undersells what we ship.

DIMENSION 1 · CARDINALITY

One agent → twelve specialists

"AI Agent" sounds singular. Mythal is twelve named agents with typed contracts and signed handoffs. The depth of reasoning at each step is what produces auditable decisions — not a single chatbot prompt.

DIMENSION 2 · TOPOLOGY

Linear workflow → closed-loop fabric

"Workflow" implies a one-way relay. Mythal closes the loop with a Verifier and a Compliance Reporter, and a Supervisor that drives a finite-state machine per finding. The trace is the audit log.

DIMENSION 3 · SCOPE

IT update → IT + OT-aware response

"Update the system" is an IT verb. In their environment, half the assets that matter are OT — and the right "update" is sometimes a compensating control + scheduled window. Mythal ships with that distinction built into a named agent.

Strategic move: say "yes — Mythal is exactly the middleware AI agent you described" in sentence one of our reply. Then turn the page and show what they didn't realize they were asking for.
07 / 12WHO ELSE THEY'RE TALKING TO

"All our vendors" — what they'll hear from the others.

Every vendor will claim to do this. Our job: know what each one actually offers so we can position by negation without naming them directly in the reply.

Qualys (VMDR + Agent Val)
Auto-remediation bolted onto Qualys VMDR. IT endpoints, mostly Windows.
Scanner-bolted (not a fabric). No OT safety model.
Tenable (Tenable One + OT Security)
Unified exposure management. OT Security product for industrial.
Workflow + visibility. No closed-loop remediation. No dedicated OT veto agent.
Rapid7 (InsightVM + InsightConnect)
Scanner + SOAR-style automation playbooks.
Playbooks are scripts, not agents. Limited reasoning depth.
IBM Autonomous Security
IT auto-remediation workflows. Watson integration.
IT focus. Compliance secondary. No dedicated OT model.
Cogent / Maze (RBVM)
Risk-based vulnerability management. Workflow + prioritization.
Workflow layer. Doesn't close the loop. No execution / verification.
ArmorCode
Application security posture management with remediation orchestration.
AppSec-shaped. Light on infrastructure / OT.
ServiceNow Security Operations
Workflow + ticketing in the existing ServiceNow stack.
Familiar to ops teams — but glue without the brains. No reasoning layer.
Splunk SOAR · Cortex XSOAR
Generic security orchestration playbooks.
Script-based automation. No agentic reasoning. No OT specialization.
Forescout · Claroty · Nozomi · Dragos
OT visibility and detection. Strong on asset inventory.
See-only. Don't close the remediation loop. We orchestrate above them.
08 / 12★ THE THREE CLAIMS WE LAND

Three sentences that beat every other reply they'll receive.

When the leadership team reads vendor responses side by side, these three claims — phrased in their language — are the differentiators.

CLAIM 1 · CARDINALITY

"Twelve coordinated agents, not one."

Each specialist owns one step of your workflow. Each handoff is typed and HMAC-signed. The reasoning trace is the audit log. No other vendor will offer this — they ship single agents or script-based playbooks.

CLAIM 2 · OT SAFETY

"Your OT lead signs the contract because we built the veto."

The OT Safety Officer agent defaults to refusing direct patches on OT-zone or CCS assets. Proposes compensating controls. Schedules firmware updates into your planned windows with dual approval. No other reply will mention this.

CLAIM 3 · COMPLIANCE

"Audit evidence is generated, not assembled."

Every closed plan emits evidence units tagged to TSA SD 1580 / NIST CSF / NIST 800-82r3 / IEC 62443. Auditor-ready PDF in under 60 seconds. Insurance attestation packet built from the same artifacts. No other vendor has this as a default output.

The opening line of our reply: "Mythal is exactly the middleware AI Agent you described — and built for the post-Mythos world by name. Here's what the fabric looks like when we map your four steps onto it."
09 / 12A CONCRETE SCENARIO

A critical CVE at 06:00 — eighteen minutes, not twenty-two days.

What happens between trigger and closed-loop verification when Mythal handles their four-step workflow on an IT asset. Demonstrable on the live console in the demo.

06:00:00

Trigger received (step 1)

Scanner Liaison ingests CVE-2026-XXXXX on corp-hr-fileserver-014. Threat Intel attaches: KEV=true · EPSS=0.81 · exploit_in_wild=true · ransomware-associated.

06:00:06

Threat level assessed (step 2)

Impact Analyst: 4,000 downstream users · regulated data · medium criticality. Change Risk: historical failure 0.03 · canary peer available · window opens 02:00 nightly. OT Safety Officer: not OT — passthrough.

06:00:10

Patch located + plan built (step 3)

Patch Hunter: Microsoft KB-5052XXX · PatchReliabilityScore=0.94. Planner builds 6-step plan with tested rollback. Policy Gate: rule SG-POL-004 → single-approval (security).

06:15:23

Approval (human gate)

Security on-call approves via console. Plan moves PLANNED → APPROVED. Executor begins immediately.

06:18:45

Applied + verified (step 4)

Executor pushes via SCCM. Verifier: rescan clean · health probe HTTP 200 · exploit retest blocked. State CLOSED.

06:18:46

Evidence emitted (their gap)

Compliance Reporter emits 3 units: NIST CSF RS.MI-01 · TSA SD §3.D · IEC 62443-2-3. Reasoning trace signed and frozen in ledger.

Total: 18 minutes 46 seconds. Legacy baseline at Class I scale for this same finding: 22 days. The Mean Time to Remediate curve is the headline KPI in the reply.

10 / 12HOW TO STRUCTURE THE REPLY

The shape of our written response.

Six paragraphs. Read in five minutes. Lands the three claims while honoring their framing.

¶ 1 · ACKNOWLEDGE

Adopt their language verbatim

Open with: "Mythal is exactly the Middleware AI Agent you described." Use their four-step workflow as the section structure. Their words become the index.

"In the wake of the Mythos release" — quote them back to themselves.

¶ 2 · MAP

Show their four steps → our pipeline

Slide 04 of this deck, in prose. One paragraph per step. Name the agents at each stage. Note the dedup, the reliability score, the rollback, the verification.

¶ 3 · EXTEND

Introduce OT without alarm

"Within thirty days of going live you'll encounter your first OT-zone finding. Here is what Mythal does that the IT-shaped workflow you described doesn't yet account for." Briefly: the OT Safety Officer.

¶ 4 · DIFFERENTIATE

The three claims

Twelve agents (not one). OT veto path (not in any other reply). Compliance evidence as byproduct (not a separate workstream). Lift these directly from slide 08.

¶ 5 · OFFER

The 90-day pilot

One IT segment + one OT zone. Engineering at-cost. Joint go / no-go at day 90. Reference pricing post-pilot: $750K–$3M ACV. On-prem appliance available.

¶ 6 · CLOSE

The two yeses + one introduction

(1) Yes to the pilot scope. (2) Yes to the timeline. (3) Introduction to their auditor + cyber-insurance broker. Frame the auditor + broker introductions as evidence the platform will hold up — not as a sales hook.

11 / 12QUESTIONS BACK TO THEM

What to ask in return — to deepen the engagement.

Embed these in the reply as "to scope the pilot precisely, can you share the following." Each answer is also intelligence we use to calibrate the demo.

Scope

  • Is the workflow scoped to IT only, or does it include OT / industrial systems?
  • What's the asset count by environment (IT / OT / DMZ / cloud)?
  • Are there Critical Cyber Systems (TSA SD 1580 scope) in play?
  • Which business units are sponsoring this initiative?

Current stack

  • Which scanners are in production today? (Qualys / Tenable / Wiz / Defender / Claroty / Nozomi / Dragos)
  • Which patch tools? (SCCM / Tanium / BigFix / Ansible / Panorama / OT-native)
  • CMDB? (ServiceNow / Device42 / iServer)
  • ITSM and SIEM in play?

Process & people

  • What's the current Mean Time to Remediate baseline?
  • Maintenance window cadence — IT and OT separately?
  • Who are the approvers? Single chain or multi-team (security + OT ops)?
  • Is there a Change Advisory Board, and how often does it meet?

Compliance & outcome

  • Which frameworks are in scope? (TSA SD 1580 / NIST CSF / NIST 800-82r3 / IEC 62443 / SOX / PCI / HIPAA)
  • Audit cycle & next major review date?
  • Cyber-insurance renewal timeline?
  • Vendor selection timeline — paper RFI now, pilot budget when?
12 / 12THE NAME · WHY MYTHAL

Why we are called Mythal — and what that signals.

Their inquiry quotes the "release of the Mythos" as the inflection point that triggered this evaluation. Our name turns that thesis into a brand asset.

ETYMOLOGY

Direct riff on the Mythos thesis

Two syllables. Easy to say. Trademarkable. The product was built post-Mythos by design, not retrofitted. Every other vendor in their RFP was named before the inflection point.

When they ask why we're called Mythal, the answer is the elevator pitch: "Because we built the platform the post-Mythos world demands."

THE MOAT IN ONE PARAGRAPH

What no other vendor will say

Twelve coordinated specialist agents with HMAC-signed handoffs. Dedicated OT Safety Officer agent holding veto rights. Closed-loop fabric with a deterministic policy gate. Compliance evidence as a byproduct. Inventory Insights agent surfacing risk before any CVE drops. Built natively for the post-Mythos operating condition.

"They asked for a middleware AI agent. We built the fabric the post-Mythos world demands."
— the one sentence that closes the deal · use it as the subject line of the reply email if you want it to be remembered.