For the Security Analyst · SOC

Triage at the speed of paste.

Bolt is the AI SuperAgent for work. Every indicator that lands in your queue (a URL, an IP, a hash, a token) is one keystroke from safe, identified, and ready to ticket. No CyberChef tab. No defang website. No paste into a model you don’t trust.

View this as a deck All roles A day in your shift, in 90 seconds

On this page

The crux

An alert fires. You paste the indicator. Bolt already knows what it is (hash, CVE, IP, JWT), defangs it, decodes it, identifies it on-device, and the background watches you set up are already running. One keystroke instead of six tabs.

The shift

Six tabs to triage one IOC. Or one keystroke.

Triage today

A tab for every step

  • VirusTotal + a defang site to make the URL safe to paste
  • CyberChef to base64 / hex / JWT decode the payload
  • A hash-ID site to name the digest, a separate tab for the CVE
  • Manual cert checks and breach lookups, when you remember to run them
  • Sensitive IOCs pasted into a consumer model you can’t vouch for

With Bolt

One SuperAgent, on-device

  • Paste the indicator. It’s recognized and defanged instantly
  • base64 / hex / JWT decoded locally, nothing leaves the box
  • Hash and CVE named on sight; flip the hash’s “Read as”
  • :cert:kev:crt:hibp watches run in the background for free
  • When you do ask AI, the IOC is masked at the boundary

Make an indicator safe to paste

Defang a URL or IP before it leaves your hands

Drop a live malicious URL or IP and Bolt hands you the defanged form (no auto-clicking, no accidental detonation), ready to paste into the incident ticket or the email to the team. Need it back? =refang reverses it.

defang ioc offline

Name a digest on sight

Recognize a hash the moment you paste it

A bare hex blob in a log line is named on sight (Bolt recognizes it as MD5, SHA-1, SHA-256 or SHA-512 by length, no sigil needed), and the “Read as” switcher lets you flip a 32-hex value between MD5 and a compact UUID when you’re not sure which it is. Need it explicit? =hashid spells out the candidate types.

hash hashid switcher

Decode a token without trusting a website

Decode a JWT locally. Never paste it on a site

Paste a JSON Web Token from a captured request (no sigil, Bolt recognizes the shape), and it decodes the header and claims right in front of you, so you can read the alg and subject without pasting a credential into jwt.io or any other site.

jwt decode offline

Deobfuscate a payload on sight

base64 / hex decode an obfuscated payload

A base64 or hex blob from a phishing lure or a script is recognized on sight (no sigil) and decodes in place, revealing the URL or command hiding inside. Chain it: decode, then =defang the result before it goes in the ticket. No CyberChef round-trip.

base64 hex deobfuscate

Recognize a vulnerability ID

Spot a CVE on sight. Tab to the advisory

A CVE id in an alert or a vendor email is recognized instantly as a vulnerability reference (typed bare, no sigil), and Tab opens the advisory page, so you stop copying it into a search bar just to find out what it is.

cve recognize link

Read network indicators on sight

IPv4, IPv6, CIDR, ASN, MAC: named the instant you paste

Every network shape that lands in your queue (an IPv4 or IPv6, a CIDR range, an ASN, a MAC address) is recognized and classified on sight, typed bare with no sigil. Private vs public, how many hosts a block covers, whose AS it is. The cheat sheet you used to keep is built in.

ipv4 ipv6 cidr asn mac

Catch a leaked secret

Spot, and redact, a leaked secret in seconds

Run =secret over a blob from a leaked config or a pastebin and Bolt names the secret it contains (Stripe key, AWS key, token) and flags it sensitive so you can hand off a masked, shareable version instead of the live credential. =pem inspects a leaked certificate or key the same way.

secret pem redact

Background coverage, for free

Set a watch once. It runs in the background

Set a watch once with the `:` sigil and a native checker runs it in the background: TLS cert expiry (:cert), CISA KEV additions (:kev), new certificates in the transparency log (:crt), breach exposure (:hibp), domain expiry (:domain), file changes (:file). They surface in Bolt when something moves, so the slow-burn signals stop slipping through.

watches cert kev crt hibp

Ask, with receipts

Ask a security question. Get a cited answer

When you do need to reach out, =ask runs a real web search and gives you a short answer with numbered [1][2] citations you can open, so you can sanity-check a claim or get context on a fresh threat without leaving Bolt, and without taking the answer on faith.

ask cited verify

What changes on your shift

Indicators triaged faster. Tools collapsed. Watches free.

1

keystroke to defang, decode, identify

recognized on-device the instant you paste
6+

browser tabs collapsed into one surface

defang site · CyberChef · hash-ID · CVE · whois · cert check
6

native security watches running for free

:cert:kev:crt:hibp:domain:file · on your machine

Why this is safe to live in

You handle the most sensitive IOCs. It stays on-device.

Defang, decode, hash-ID and secret detection run locally, with zero network and zero tokens. The indicator you’re triaging never leaves your machine just to be understood. The watches (:cert:kev:crt:hibp) query their sources on your behalf, from your machine, never through Sparcle.

On-device by default

Recognition, decoding and defanging are pure local logic: no egress, no model, no waiting. The fastest answer is the one that never left the box.

Masked when you do ask AI

If you escalate an IOC to AI, sensitive values are tokenized before the model sees them and restored in the answer, automatically, every time.

Safe while you triage

Every AI action is masked on-device before any model, so the analyst’s speed never becomes the security team’s exposure.

Governed AI

Detected sensitive values are tokenized before the prompt leaves your boundary

Use AI on real work without leaking the identifiers in it. Emails, card and account numbers, SSNs, phone numbers and keys are masked before anything reaches the model and restored in the reply, and when you point it at your own LLM, the rest stays inside your boundary too.

privacy governed safe

Triage, not exposure

Use AI on the alert without feeding it the evidence.

Indicators, hostnames and user identifiers are masked before the model call, and the retrievals, tool calls and model calls land in a chain you can verify offline.

New: the governed browser

The browser Bolt ships, and the check on the way out

Most AI exposure happens in a browser tab. Bolt ships its own, so what a person types or pastes into a page is screened on the machine before it is allowed to leave it.

A normal browser

Sees the paste, says nothing

  • A customer list reaches the AI site intact
  • The extension is the only control, and only where it is installed
  • No record of what was sent, or why
  • Every AI-enabled extension holds its own permissions

Bolt's browser

Names it, holds it, records it

  • What was found, grouped by type with a count, without repeating the values back
  • The page is held while the person decides, and a written reason is required to proceed
  • The reason lands in the audit record as a durable incident, not a counter
  • The tab says Guarded only when the guard reported, and Not screened when it cannot

The ask

Put it on every SOC desk. Triage gets faster today.

Bolt is free to roll out across the team: defang, decode, hash-ID, CVE recognition and the native watches all work on day one. Every IOC you escalate to AI is masked before it ever reaches a model.

Free

the SuperAgent on every analyst’s desk

$0

tokens for on-device triage

1

keystroke from paste to ticket

Every other seat

Same platform, argued from where you sit

Each role page makes the case in that role's own language, and links back here.

See it on your own data

Bolt runs inside your perimeter, on the LLM you choose.