Skip to content
Hosanna Tecnologia Hosanna Tecnologia Talk to us

Security

Discovery

What has already leaked from your company, what is exposed to the internet, and whether it still opens the door

What has already leaked from your company, what is exposed to the internet, and whether it still opens the door.

Every company has credentials that have already leaked through a vendor, published services nobody remembers turning on, and lookalike domains registered by third parties. Discovery measures this exposure without sending a single packet to your infrastructure, delivers a diagnosis that fits in four pages, and leaves the one question that requires privileged access for you to answer, on your own premises.

  • Collection that never touches your network
  • Nothing gets in without your authorization
  • Provenance for every claim
  • Designed to ISO/IEC 27001 and 27701
  • 0 packets sent to your infrastructure during collection
  • 4 pages in the executive report, a design limit and not a coincidence
  • 3 versions of the report, one for each audience that will read it
  • 1 check enters your network, and it runs with your administrator's credential

01 · The challenge

Exposure that has already happened, and that nobody has measured

The question is not whether your company will be a target. It is how much of it is already published, and for how long. Three situations come up in practically every assessment, and none of them depends on anyone having broken into your systems.

  • The credentials are already out

    Your company was not breached. A service your team used was, and the corporate password was in it. From that day on, the credential circulates in databases anyone can search, and your security team is not notified.

  • The surface grew without an inventory

    A service published for a test and never shut down. A certificate issued for an address nobody recognizes. An old version the internet can read and that already has a publicly known flaw. None of this shows up in the internal inventory.

  • The question left unanswered

    Suppose the list of leaked credentials exists. The next question is the one that decides the budget: do any of them still work today? Without that answer, the survey is interesting but not actionable.

Why this became a board-level issue

A reused credential is the cheapest way in there is, and it does not require any technical flaw on your side. As long as the security conversation stays with the people who run the network, the decision to invest keeps being made without the number that justifies it.

02 · What Discovery is

An exposure diagnosis you can take to the board

Discovery is a defensive tool. It exists so that an organization can understand its own exposure, and it was built with controls that prevent its use as a reconnaissance instrument against third parties.

The three questions

  1. Which corporate credentials are exposed. Addresses on your domain that appear in leaks and in stolen credential databases, with the date they were seen, the category of data involved and, when the source provides it, the password itself.
  2. What surface is visible from outside. Published services, known flaws associated with the exposed versions, names issued in certificates, the domain's registration status and the hygiene of email sent in its name.
  3. Whether those credentials still work. This is the one with the greatest impact in the boardroom, and the only one that requires privileged access. It is answered inside your perimeter, by your team, with its own credential.
  • What it is

    • A diagnosis with a reference date, a stated validity period and provenance on every line.
    • A conversation that starts with a number and ends with an action plan ordered by effort.
    • A document written on the assumption that your technical team will challenge every finding.
  • What it is not

    • It is not a penetration test, and it does not replace one.
    • It is not continuous monitoring. It is a dated snapshot, and the document says so on its own cover.
    • It is not a dashboard left running and consuming queries without anyone having asked for a measurement.

03 · What it delivers

Six fronts, a single picture

Each front answers a question your company already has, and none of them requires access to your network to be answered.

  • Exposed credentials

    Identities on your domain found in leaks, with the observation date, the compromised data categories and how many include a password. The exposure rate comes from here.

  • Attack surface

    Services the internet can see, recognizable versions and the public flaws associated with them, sorted by severity.

  • Known exploitation

    A cross-check against the public catalog of flaws under active exploitation. This is the line that separates a theoretical vulnerability from one that is already being used, and it is the strongest argument for prioritization.

  • Supply chain

    Vendors identified from your domain's public configuration, which of them have a known leak, and which are authorized to send email on behalf of your company.

  • Domain and email hygiene

    Registration status and expiration, ownership, names issued in public certificates, and the configuration that protects your domain against misuse in messages.

  • Domain portfolio

    For organizations with more than one brand or more than one company in the group, a view of the whole set, with its own document and the same identity protection rules.

04 · How it works

Two sides, and a boundary nobody crosses over a connection

The product has two sides with different trust boundaries, and they do not talk over a network. What passes from one to the other is a file, signed and then encrypted, which you deliver through whatever channel you prefer.

Hosanna's operation

Public and third-party databases

leaks, exposed services, certificates, registration, catalogs

reading, never scanning

Analysis and report

consolidates, scores, writes and issues the signed authorization

signed and encrypted authorization signed and encrypted verdict

Your perimeter

Your administrator's workstation

a separate application, opened by them, with no internet access

Your corporate directory

read with their credential, locally

Neither side opens a listening port: the absence of a network surface is a design decision, verified with every release, and not a side effect of how the product turned out

The exchange is a file, not access. Signing and then encrypting are two different guarantees, and both are present: encryption keeps third parties from reading the content, and the signature proves who issued it. Encrypting without signing would let anyone produce a file that your workstation would agree to open.
  • Encryption

    X25519 key exchange with ChaCha20-Poly1305 authenticated encryption, using an open and audited library. The product does not build its own cryptographic scheme.

  • Signature

    An Ed25519 signature over the content, checked on arrival. A file tampered with to point to another target is rejected before anything runs.

  • Scope

    The authorization carries the scope you validated, and your administrator's workstation refuses to run outside it. The scope is not a recommendation. It is a condition for running.

05 · What can be seen from outside

Collection that never touches your network

The entire first phase consists of reading databases that have already done the collecting, plus public domain and certificate lookups. At no point does the product send scanning traffic against your infrastructure, and the report states this on a page of its own.

What is read, and what each source answers
FrontWhat is queriedWhat you get
CredentialsPublic and commercial leak databases, by address and by domainAffected identities, observation date, data categories and how many include a password
Exposed servicesPublic inventories of services reachable from the internetVisible ports and services, recognizable versions and associated flaws
Active exploitationPublic catalog of flaws under known exploitation, downloaded in full and cross-checked locallyWhich of the flaws found are already being exploited in the real world
Names and certificatesPublic certificate transparency logsAddresses issued in your domain's name, including the ones you do not recognize
Domain registrationOfficial registration lookup serviceRegistrant, creation date, expiration and status
Chain and emailPublic domain and message sending configurationInferred vendors and who is authorized to send email on your behalf

The cross-check against the exploitation catalog is done on our side

The entire catalog is downloaded and compared locally. Querying flaw by flaw would reveal your company's inventory to the source being queried, and that is exactly what a diagnostic tool must not do.

Every finding says how far you can trust it

A confirmed finding and an inferred finding are not the same thing, and the report never presents them with the same weight. A flaw deduced from the version a service announces is written in the conditional and marked as such, because it needs internal confirmation to become a certainty. The absence of a finding is not proof of the absence of exposure either, and the document says so in writing instead of leaving it implied.

Opening the dashboard does not trigger a measurement

Viewing the result of an assessment reads what has already been collected. A new query to the external sources only happens when someone asks for it, because measuring has a cost and no one should receive that bill without having decided to pay it.

06 · The internal check

Do those passwords still open the door?

This is the question with the greatest executive impact and the greatest operational sensitivity. That is why we do not answer it: your team answers it, on your network, with its own credential, and only the result comes back.

All of this happens inside your perimeter

Authorization

signed and encrypted, delivered to you

Your administrator

opens it, checks the scope and consents

The directory is read

with their credential, through the standard mechanism

Comparison

in memory, on site

Verdict

yes or no per identity

What never leaves your network: no cryptographic hash of your directory, no password and no credential material of any kind

The comparison proves without trying to get in. The product never tries to authenticate with any credential as a verification method. Comparing cryptographic hashes is the only authorized method, and that is what separates a verification from an intrusion attempt.
  • The part that touches credentials is small on purpose

    The part of the product that handles credential material is kept minimal, cohesive and isolated, precisely so that your security team can audit it on its own, without reviewing the whole product. Functionality that does not need to be there stays out.

  • No internet access, and no key of ours

    The application that runs on your premises works without any external connection, and it carries no access key to any vendor. Everything it needs arrives sealed inside the authorization you approved.

07 · The report

The deliverable is the document, and it was designed for the room

Collection, scoring and cryptography exist to produce the report. A technically flawless diagnosis whose document cannot carry a meeting is a wasted diagnosis, and the document is read by people with incompatible profiles in the same room.

Three versions from the same assessment, one for each reader
VersionFor whomIdentitiesLength limit
BriefWhoever approves someone else's decision, in ninety secondsAlways aggregated1 page
ExecutiveExecutive leadership and the boardAlways aggregated4 pages
TechnicalSecurity and technology, to verify and actNamed, with raw evidence in an appendixno limit

The brief is not a summary of the executive version. The difference is the reader: the executive version is read by whoever will decide the budget, the brief by whoever will approve someone else's decision, almost always on a phone, between two meetings. That is why it opens with what changed, not with the number.

The four pages of the executive version

  1. Exposure summary. Customer, domain, reference date and validity. The exposure score with a named band and, as the most prominent number on the page, the exposure rate: how many corporate identities are compromised, out of how many known. Between three and five of the most severe findings, in business language, and a one-sentence factual verdict.
  2. Where the risk is. The exposure broken down into three blocks with absolute numbers: credentials, attack surface and supply chain.
  3. Prioritized action plan. A table ordered by the ratio between impact and effort, in which each action states how much the score drops if it is carried out, the effort involved and the suggested owner.
  4. Methodology and limitations. Mandatory, and never omitted. Which sources were queried, that the collection was exclusively passive and made no contact with your infrastructure, that some findings are inferred and call for internal confirmation, and that the absence of a finding does not prove the absence of exposure.

The column that turns a report into an approved budget

Stating how much the score drops with each action is what turns a diagnosis into a work plan. That column is how the security team justifies its priorities, and it is how leadership can compare two actions without having to understand the technique behind either one.

The six writing rules of the document

  • Mandatory provenance

    Every statement can be traced to evidence, a source and an observation date. A statement without provenance does not go in.

  • No alarmism

    No fear-driven language, dramatic superlatives, countdowns or unfounded loss estimates. The real exposure is persuasive enough.

  • Explicit confidence

    Each finding shows how far it can be trusted, and what is inferred is written in the conditional.

  • Always with a way forward

    Every finding comes with an actionable recommendation, an estimated effort and a suggested owner. A document that only raises alarms reads like a sales pitch.

  • Holds up under challenge

    The text is written on the assumption that your team will challenge every line, and the raw evidence is available for independent verification.

  • Classified on every page

    The document carries the confidentiality marking on every page, and it closes by stating what was not looked at.

08 · Personal data and authorization

Proof comes before the name

A diagnosis of exposed credentials handles personal data by nature. The product deals with this by reversing the usual order: the finding appears first, and the identity only after someone proves that the domain is yours.

The domain is queried

the diagnosis appears with identities obfuscated

Proof of ownership

a record published in your domain's DNS

Reveal becomes available

and every reveal becomes an audit record

Authorization

only for a domain proven to be yours

The proof enables the reveal action, not the content: the difference between saying “I can't show you what I found” and saying “I found this, and to see whose it is you need to prove the domain is yours”

The check happens where the output is assembled. The choice to reveal applies to the session, but the proof applies to a domain. That is why the check is made when the screen, the spreadsheet and the document are assembled, and not at the moment someone chooses to reveal. A product that checked only at the moment of choice would let the identities of the next domain come out in full.
  • The password, when the source provides it

    Some sources return the password itself. It is stored sealed, displayed masked by default, and only appears in full when both conditions hold at the same time: proof of ownership of the domain and an explicit choice in the session. Showing the beginning and the end is what lets the owner recognize their own password and validate the diagnosis at the table.

  • Where it never appears

    In the delivered document, in the exchange file and in the audit trail. These are the three places that leave the product and that nobody watches afterwards, so the ban there is absolute, regardless of any permission.

One key per assessment

Each assessment has its own data key, and the stored data is encrypted at rest. Closing an assessment means destroying its key, which makes the content unrecoverable without depending on deleting record by record.

09 · What it does not do

The limits are written down, and they are verified

A tool that measures other people's exposure is a tool that can be used against third parties. The restrictions below are not internal policy: they are part of the product's design and are checked with every release.

  • Does not exploit flaws

    No vulnerability exploitation, no lateral movement and no post-compromise action, under any circumstances.

  • Does not try passwords

    No authentication attempt against any service, internal or external. Checking reuse by comparison is the only authorized method.

  • Does not scan without authorization

    No active scanning against a domain or infrastructure without verified authorization. A domain that has not been proven receives no authorization at all.

  • Does not build a database of everyone

    A store that combined the exposure of several customers would be a multi-organization attack map. It is not built, and the decision not to build it is recorded in the design.

What stays out, and rightly so

Multi-factor authentication posture cannot be observed from outside. From outside, it is only possible to identify which identity provider the company uses, and any claim beyond that would be a guess presented as a finding. The product says what it sees and where it cannot see.

10 · ISO and LGPD

Controls in the design, evidence in the process

The product was designed according to the requirements of the standards below, and what it produces serves as evidence for anyone who needs to demonstrate compliance.

  • ISO/IEC 27001

    Information classification on every generated document, encryption at rest per assessment, separation of trust perimeters, an access trail and control over who can reveal what.

  • ISO/IEC 27701

    Personal data obfuscated by default, proof of legitimacy before any reveal, and an auditable authorization trail in which revealing is a recorded event and not an implicit state.

  • Brazil's General Data Protection Law (LGPD)

    Stated purpose, minimization at the source, retention tied to the assessment and disposal by key destruction. The delivered report never carries a credential in plain text.

A note on the artificial intelligence standard

ISO/IEC 42001 covers artificial intelligence management systems. It does not apply to this product today, and we say so here instead of leaving it out. The certifications listed at the end of this page refer to Hosanna Tecnologia and to the scope of its certified systems.

11 · When to use it

Six situations where the diagnosis decides something

  • Before approving the security budget

    The conversation stops being about abstract risk and becomes about a dated number and a plan with estimated effort. Leadership compares actions, not opinions.

  • During due diligence

    Buying a company means inheriting its exposure. The diagnosis measures it without asking for access to the other party's infrastructure, which is exactly what an ongoing negotiation allows.

  • After a vendor's incident

    When a service your team used is compromised, the immediate question is how many people from your company were in it and whether any of them reused the corporate password.

  • When preparing for an audit

    The document provides dated evidence with provenance, and the authorization trail shows how personal data was handled during the assessment itself.

  • In periodic reviews

    Repeating the assessment turns the snapshot into a trend, and comparing two dates shows what your team actually closed in the period.

  • In a group with several brands

    Each company gets its own diagnosis, and the holding company gets a view of the whole, with the same identity protection rules applied.

12 · How the engagement works

From scope to report, with a decision of yours in the middle

The engagement has seven stages, and the fifth is a decision point only you can make. Up to that point, nothing has touched your infrastructure.

  1. 1 · Scope and domains
  2. 2 · Proof of ownership
  3. 3 · No-contact collection
  4. 4 · Joint review
  5. 5 · Your authorization
  6. 6 · Check on your premises
  7. 7 · Consolidated report

customer decision point

Four reference weeks, and the fifth stage is yours. Timelines vary with the number of domains and with your team's availability for the joint review. What does not vary is the order: the internal check never happens before you have seen the diagnosis and authorized the scope.
  1. Scope and domains

    We define which domains are included and what the reference date is. The scope is written down before any query.

    You get: the agreed scope in writing.

  2. Proof of ownership

    Your team publishes a record in the DNS of each domain. That record is what authorizes revealing identities and what opens the way to stage 5.

    You get: the exact instructions for the record to publish.

  3. No-contact collection

    Reading the public and third-party sources happens entirely on our side. No packet reaches your infrastructure.

    You get: nothing yet, and that is intentional.

  4. Joint review

    We present the diagnosis to your team before any internal check, so they can challenge whatever they want to challenge.

    You get: the executive and technical versions.

  5. Your authorization

    You decide whether you want to answer the third question, and with what scope. If the answer is no, the engagement ends here with value delivered.

    You get: the signed authorization, to open whenever you want.

  6. Check on your premises

    Your administrator runs it, with their own credential, inside your perimeter. We can follow along on a call if you wish, with no access at all.

    You get: the result file, which you keep as well.

  7. Consolidated report

    The final document brings together what is seen from outside and what was confirmed inside, with the prioritized action plan.

    You get: the three versions of the report.

13 · What your company gains

The same assessment, read from four different seats

  • For the CEO and the board

    A dated number and a plan with effort. The question “are we exposed?” gets a one-sentence answer, and the decision to invest no longer depends on who argues best.

  • For information security

    Findings with provenance and raw evidence for independent verification, prioritization by real impact, and confirmation of reuse without opening the network to any outsider.

  • For IT

    An inventory of what the internet can see, including what nobody remembered publishing, and the list of names issued in certificates that are missing from the internal inventory.

  • For privacy and compliance

    Dated evidence for audits, an authorization trail for the processing done during the assessment itself, and a document that never carries a credential in plain text.

Indicators to track between two assessments

  • Exposure rate: compromised identities over known identities
  • Exposed flaws that are under known exploitation
  • Visible services that are missing from the internal inventory
  • Leaked credentials that still worked in the directory

14 · Frequently asked questions

Straight answers to the most common questions

Are you going to scan our network?

No. Collection consists entirely of reading databases that have already done the collecting, plus public domain and certificate lookups. No packet goes from us to your infrastructure, and the report states this on one of its four pages.

Then how do you confirm whether the password still works?

We don't. You do. The check runs in a separate application, inside your perimeter, opened by your administrator and with their credential. The comparison happens in memory, on site, and only the verdict per identity comes back.

Do you keep any passwords or any data from our directory?

No cryptographic hash of your directory and no credential material leaves your network. The passwords that appear in the diagnosis come from public leaks, are kept sealed, are displayed masked and never go into the delivered document, the exchange file or the audit trail.

Do we need to install anything?

Only if you decide to run the internal check, and only on your administrator's workstation. It is a desktop application that opens no listening port, runs without internet access and carries no access key to any vendor.

Why do you ask us to publish a record in our DNS?

Because that is what proves the domain is yours. Before that proof, the diagnosis appears with every identity obfuscated: you see which leak, when it happened and what to do, but not whose it is. The proof enables the reveal action, and every reveal is recorded.

Does this replace a penetration test?

No, and the two answer different questions. A penetration test asks whether someone can get in by exploiting flaws. This diagnosis asks what has already leaked and what is already visible, without touching your infrastructure. It usually comes first, because it costs less and tends to find the cheapest way in.

Is the report useful for audits?

It serves as dated evidence with provenance. Each statement can be traced to a source and an observation date, the technical version includes the raw evidence in an appendix, and the document explicitly states what was not looked at.

What if you don't find anything?

The document will say so, and it will also say that the absence of a finding does not prove the absence of exposure. A report that presents itself as exhaustive falls apart at the first hard question, and stating the limitation is what keeps the rest credible.

15 · Next steps

Start with one domain

The shortest conversation starts with your main domain and a review meeting. Up to stage four, nothing needs to be installed and nothing touches your infrastructure.

  1. Choose a domain. The main one is usually the most revealing, because that is where the corporate addresses are.
  2. Publish the proof record. It takes your network team a few minutes, and it is what makes it possible to talk about identities later.
  3. Receive the diagnosis and challenge it. The joint review exists so your team can question the findings before any decision.
  4. Decide on the internal check. If it makes sense, you authorize the scope and run it on your premises.

16 · Contact

Talk to Hosanna Tecnologia

Let's talk about your domain, your situation and what makes sense to measure first.

Certifications

Discovery uses technology developed and maintained by Hosanna Tecnologia Ltda; the certifications refer to Hosanna Tecnologia Ltda and to the scope of its certified systems.

Back to top