# Whisper Endpoint - the EDR that governs identity and holds the off-host kill-switch

endpoint.whisper.online. A real Linux EDR with a difference no host-only sensor
has: it gives every host and every AI agent a routable, publicly-verifiable IPv6
/128 on our own address space, and it can end a compromised host or agent at the
network layer even when the host is fully owned. Written so an agent or a person
can act on it.

## The question

You can summon an autonomous agent in seconds. Can you kill it if it turns? A
host-only EDR runs inside the blast radius: when an attacker (or your own agent)
owns the host, it owns the sensor, and the one control that matters, ending the
thing, is exactly the control it can no longer reach. Whisper Endpoint closes that
gap by governing identity and territory, and it is also a real sensor. Here is the
receipt.

## A live, verifiable identity (verify it yourself, keyless)

Every host and agent gets a routable IPv6 /128 out of 2a04:2a01::/32, announced by
AS219419, with DNSSEC-anchored reverse DNS, a DANE-EE key pin, and a public RDAP
registration. Verify with stock dig, against the IANA root, with our API nowhere
in the trust decision.

**Keyless** means no Whisper API key and no account: the cryptographic keys are all
still there, in DNSSEC and DANE, and checking them is the point. **Trustless** means
you do not have to believe our API or the resolver in between, because every DNS answer
is checked against the DNSSEC chain from the IANA root. It does not take the
operator of the signing zone out of the path; the independently witnessed,
tamper-evident transparency log is there for that, and under bring your own domain
the `DS` at your own registrar decides whether our signatures count.

    $ dig -x 2a04:2a01:e32a:155f:26cc:967b:8c5b:70a1 +short
    a26cc967b8c5b70a1.t0f1c527d1cada8c6e8dfc46c74e013e5.agents.whisper.online.

## The receipt (measured, not claimed)

We drove real Atomic Red Team atomics against the shipped Linux sensor on a
throwaway Debian 12 box (kernel 6.1) and measured what it captured. No mockups.

- 100% telemetry: 37 of 37 ATT&CK techniques captured across 10 tactics, zero drops.
- 32 of 37 signature (86%): the known tool captured by name (nmap really scanned,
  gpg really encrypted, /etc/ld.so.preload write captured). The five not named by
  signature were still captured as telemetry.
- 17,092 exec events captured, zero drops, including one scan loop of 10,974.
- Scope: the Atomic Red Team Linux universe is 113 techniques / 483 atomics; this
  curated cross-tactic run exercised 37 techniques in 10 tactics.

Verbatim captured events (the screenshots, per our prove-then-document rule):

    [T1003.008 read /etc/shadow]  {"comm":"cat","filename":"/usr/bin/cat","pid":1658,"ppid":1656,"uid":0}
    [T1489 service stop]          {"comm":"systemctl","filename":"/usr/bin/systemctl","pid":1718,"uid":0}
    [T1053.003 cron persistence]  {"comm":"crontab","filename":"/usr/bin/crontab","pid":1732,"uid":0}
    [T1547.006 kernel module]     {"comm":"insmod","filename":"/usr/bin/kmod","pid":1875,"uid":0}
    [T1027 obfuscation]           {"comm":"base64","filename":"/usr/bin/base64","pid":1986,"uid":0}
    [T1071.001 C2 user-agent]     {"comm":"curl","filename":"/usr/bin/curl","pid":2018,"uid":0}
    [T1485 data destruction]      {"comm":"dd","filename":"/usr/bin/dd","pid":7462,"uid":0}
    [T1105 ingress transfer]      {"comm":"rsync","filename":"/usr/bin/rsync","pid":7495,"uid":0}
    [T1574.006 ld.so.preload]     {"kind":"file","event":"create","path":"/etc/ld.so.preload","critical":true}

## The blind spot host-only EDR cannot close

- Owned host, owned sensor. Full compromise of the host is full compromise of the
  sensor: it can be unhooked, starved, or silenced. Containment on the same host
  as the threat is a request the threat can decline.
- Whole classes it never sees. You cannot install a kernel agent on a PLC, camera,
  printer, a partner's laptop, or inside every ephemeral AI agent. Those are most
  of the fleet.

## The turn: identity, then the off-host kill-switch

- Identity: a routable /128 the internet can check with dig, DNSSEC + DANE + RDAP.
- Containment: one call revokes a compromised host or agent, its resolution and
  egress stop worldwide at DNS-TTL. The lever lives on our infrastructure, not on
  the host, so it works even when the host is fully owned. The one deliberate act
  the golem's own machine cannot veto.

## ATT&CK coverage (15 tactics, 712-technique taxonomy)

We already run the ATT&CK taxonomy in the graph (712 techniques, 15 tactics). Two
of the biggest tactics, reconnaissance and resource-development, are the attacker's
own infrastructure: a host-only sensor is structurally blind there; our graph is
not. Strong today: execution, persistence, discovery, command-and-control,
exfiltration, plus reconnaissance and resource-development (graph-only). Partial and
growing (telemetry captured, verdict layer growing): credential-access, impact,
privilege-escalation, collection, lateral-movement, initial-access, defense-evasion,
impair-defenses. Of the 37 techniques driven live, all 37 were captured.

## Coverage of the uncoverable

- OT / IoT (PLCs, cameras, printers): a routable identity + default-deny egress you
  govern from the network; one call cuts it off.
- Agents and sub-agents: each gets its own /128, its own logs, its own revoke.
- Personal and contractor devices (partners, freelancers): attribution and a
  kill-switch without pushing
  software onto someone else's laptop.

## Calm footprint (a published cgroup SLA)

- CPUQuota=25%: a hard kernel ceiling of a quarter of one core.
- MemoryMax=250M: a hard cap in the unit, the same on every host.
- Event-driven, fail-open: no busy polling; if the sensor is down, the host still
  resolves and runs. Never a self-inflicted outage.

## Compliance evidence

Every identity and control event is written to an independently-witnessed,
tamper-evident ledger an auditor checks themselves, keyless. Per-control evidence
toward SOC 2 (CC6.1 / CC6.2), ISO 27001 (A.5.16), HIPAA (164.312), PCI DSS
(Req 1 & 8), CMMC, NIS2 and DORA. It provides a control and the evidence for it; it
never makes you compliant or certified. Your auditor does that.

## No asterisks - in plain words

Telemetry is visibility: the measured 100% means the sensor saw every technique and
named its signature process (exactly what MITRE Evaluations report first). It is not
the same as an analytic verdict that alerts. Detection is the growing layer: live
for graph reputation on C2/exfil destinations and file hashes, the per-process
trust tier, and the behavioral engines (ransomware T1486, the dropper event chain
T1105, process injection T1055), observe-first on Linux and Windows today; and we
never claim a detection we cannot demo.

- Live today: the ransomware verdict fires (T1486 at confidence 1.0 in the live
  run). Copy-on-write rollback is the roadmap half, turning a held verdict into an
  undone incident.
- Live today: the Windows deep tier, ETW telemetry plus the ransomware, dropper and
  injection engines, proven on Windows Server 2022. ETW-TI operation-time injection
  visibility stays built but dark, gated on Microsoft's anti-malware attestation,
  an external clock, not a date we invent.
- 2027: a public MITRE Evaluation, so the coverage here is scored by someone else.

## Install (one command)

    curl -fsSL https://get.whisper.online | sh

One command mints the identity you can then audit with dig -x, anchored at the IANA root; add WHISPER_SENSOR=1 and a key to install the calm sensor too.

## Pages
- /               Overview - the question, the receipt, the coverage
- /how-it-works   The sensor, the identity, and the off-host kill-switch
- /setup          Install and verify, one command, then dig -x
- /privacy        EU sovereignty and data residency
- /pricing        Per-seat, no minimum, direct or through a partner

## The operator
Operator: Whisper Security (viaGraph B.V.), Amsterdam, NL. Network: AS219419,
IPv6-only, RPKI-signed, MANRS-compliant, announcing 2a04:2a01::/32. Built by an
ex-RIPE NCC CIO and a K-root architect.

## Get started
Sign in: https://console.whisper.security/sign-up?redirect_url=https%3A%2F%2Fconsole.whisper.online%2F
Reselling to a book of clients: https://www.whisper.security/solutions/mssp
Building agents or a fleet: https://whisper.online/