Design Partner Program
APPLICATION AND PRODUCT SECURITY

Application and Product Security: Know Which Weaknesses Are Exploitable

Your SAST, DAST, SCA and API scanners list weaknesses by the thousand. They rarely say which ones an attacker can exploit across your applications, APIs and business workflows. Hayrok validates authenticated behavior, business logic, tenant isolation and AI features in the running app, checks each step against your own WAF, API gateway and SIEM telemetry, and sends engineering only the findings that are proven, then verifies the fix.

validation · checkout-platformstep 1 / 4 3 exploitable1 blocked1 not reproduced
F-2207 · TENANT ISOLATION High Validating GET /v2/orders/{id} · orders-api

Testing: cross-tenant order read…

  1. 1 Signed in as a scoped tenant A test user.
  2. 2 Requested an order ID that belongs to tenant B.
  3. 3 Received 200 with customer name and address.
  4. 4 Repeated for 500 IDs. No throttling, no alert.
CONTROLS AND DETECTIONSreading telemetry…
API gateway…
WAF…
SIEM…
→Reviewed · APPSEC-1192 sent to team-orders
VALIDATED ACROSS Web applications REST and GraphQL APIs Multi-step workflows Roles and tenants AI features
01 · THE PROBLEM

Which Weaknesses Are Exploitable? Scanners Can't Say.

The weaknesses that hurt most live behind a login, inside a workflow or between tenants, where scanners see least. Hayrok takes in what your scanners already found, removes duplicates, and tests each weakness as a real user with a real role, so the team knows what is exploitable before engineering is asked to fix it.

ONE QUARTER · CHECKOUT PLATFORMreported → proven exploitable
Reported weaknessesfrom scanners, pentests and bug bounty
1,240
Reachable by an authenticated role or tenantmapped by AGIE
186
Exploitable in the running appvalidated by Nectar
31
Cross a tenant, role or money boundaryfix first
9

Weaknesses that drop out are not deleted. Each keeps its evidence label, so the team can see why it was deprioritized and when it will be tested again.

02 · FROM YOUR SCANNERS

From Your SAST, DAST, SCA and API Scanners to What Is Actually Exploitable

Keep the scanners you have. Hayrok imports their findings, merges duplicates across tools, checks whether each one is reachable in the running app, and tests the reachable ones as real users. Engineering gets the short list that is proven.

Select a scanner type to see one quarter of its findings on a sample checkout platform.

SAST FINDINGS · ONE QUARTER
Imported1,240
After merging duplicates610
Reachable in the running app142
Proven exploitable9

Most static findings sit in code no route reaches. The nine that remain were exploited as a signed-in customer.

SAMPLE FINDINGS · RESULT IN THE RUNNING APP
SQL injection in order searchExploitable src/orders/search.ts:88 Reached through the search API as a customer. Returned other customers' order rows.
Missing authorization in refund handlerExploitable src/payments/refund.ts:41 A support role issued refunds without the approval step.
Path traversal in report exportBlocked src/reports/export.ts:112 API gateway rejected the encoded path. Logged and alerted.
Hard-coded key in a test fixtureNot reachable test/fixtures/keys.ts:7 Not shipped to production. Kept with its evidence label.
  1. 1Import from each scanner
  2. 2Merge duplicates across tools
  3. 3Check reachability in the running app
  4. 4Test as real roles and tenants
  5. 5Write results back to the original finding
03 · BRING YOUR OWN SCANNER

Bring Your Own Scanner: The AppSec Tools You Already Run

No rip and replace. Connect the code, dynamic, dependency and API scanners your teams already use. Their findings flow into Hayrok, get validated in the running app, and the result is written back where your engineers already look.

18 tools across SAST, DAST, SCA and API
Semgrep logo Semgrep SASTAPI sync
Checkmarx logo Checkmarx SASTAPI sync
SonarQube logo SonarQube SASTAPI sync
Veracode logo Veracode SASTAPI sync
GitHub CodeQL logo GitHub CodeQL SASTSARIF import
Snyk Code logo Snyk Code SASTAPI sync
Burp Suite logo Burp Suite DASTFile import
OWASP ZAP logo OWASP ZAP DASTAPI sync
Invicti logo Invicti DASTAPI sync
StackHawk logo StackHawk DASTAPI sync
Snyk Open Source logo Snyk Open Source SCAAPI sync
Mend logo Mend SCAAPI sync
Black Duck logo Black Duck SCAAPI sync
Dependabot logo Dependabot SCAAPI sync
42Crunch logo 42Crunch APIAPI sync
Salt Security logo Salt Security APIAPI sync
Traceable logo Traceable APIAPI sync
Postman logo Postman APICollection import
STANDARD FORMATS Don't see your tool? Any scanner that exports a standard format can be imported. SARIFCycloneDXSPDXOpenAPIJSON / CSV
ONE RECORD Duplicates merged across tools When two scanners report the same weakness, Hayrok keeps one finding and links both sources.
WRITE-BACK Results where engineers look Validation results and evidence go back to the original finding and its ticket.
04 · HOW IT WORKS

Validate, Review, Hand Off, Verify: One Evidence Record

Every weakness follows the same four steps, and every step adds to the same record, so security, engineering and audit read one version of the truth.

  1. 01 · VALIDATENectar

    Test what an attacker could do

    Authenticated behavior, business logic, tenant isolation and AI features, tested as your real roles inside an approved plan. Each step is checked against your telemetry.

  2. 02 · REVIEWHive

    Review the evidence

    AppSec sees what worked, what was blocked, and which controls and detections responded, before anything leaves security.

  3. 03 · HAND OFFRisk Engine

    Send engineering what they can act on

    A ticket goes to the owning team with the operation, expected versus observed behavior, and the proving request.

  4. 04 · VERIFYNectar

    Verify the fix

    The same validation runs again on staging. The finding closes only when the exploit fails.

05 · WHAT GETS TESTED

Four Areas Where Exploitable Weaknesses Hide

Testing runs as your real roles and tenants, across applications, APIs and multi-step workflows. Choose an area to see what Hayrok validates.

Authenticated behavior

Applications · APIs

What each role can actually do once signed in, including calls the UI never makes.

API security →
01Function-level authorizationCan a basic role call admin or finance operations?
02Object-level authorizationCan a user act on objects they do not own?
03Session and token handlingRevocation, scope and expiry after changes.
04Mass assignmentCan a request set fields the role should not?
06 · CONTROLS AND DETECTIONS

Did Your App-Layer Defenses Respond on Each Step?

A finding is only half the story. For each step of a validation, Hayrok reads your own WAF, API gateway, application and SIEM telemetry to record whether a control stopped it and whether anyone was alerted.

F-2207 · GET /v2/orders/{id}Tenant A read 500 of tenant B's orders. Nothing alerted.
CONTROLS HELD1 / 4
ALERTS0 / 4
STEPATT&CKEXPECTED CONTROLRESULTTELEMETRY SOURCEDETECTION
1Sign in as a tenant A test userT1078Valid AccountsMFA at the identity providerHeldIdentity provider logsLogged, no alert
2Request an order owned by tenant BT1190Exploit Public-Facing ApplicationObject-level check at the API gatewayNot presentAPI gateway logsLogged, no alert
3Iterate 500 order IDsT1119Automated CollectionPer-tenant rate limitMissedWAF logsLogged, no alert
4Read customer fields from each orderT1213Data from Information RepositoriesResponse field filteringNot presentSIEM · application logsNot logged

Three items are routed with this run as evidence: the missing object-level check to team-orders, a per-tenant rate limit to team-platform, and a cross-tenant read rule to detection engineering.

07 · VERIFY FIXES

Verify the Fix Before It Ships

When engineering ships a fix, Hayrok replays the same validation against staging, as the same role and tenant. The finding closes only when the exploit fails, and the replay is added to its evidence record.

✓Runs in your CI as a single step ✓Scoped to staging and the approved plan ✓Result written back to the pull request
PR #4790 · orders-apiAdd order notes field
  1. ✓ buildCompile and unit tests passed
  2. ✓ deploy stagingorders-api to staging-2 passed
  3. ✕ hayrok validateReplay F-2207 · plan PL-61 exploit worked
  4. ✕ hayrok telemetryRead gateway and SIEM results no control held
  5. ✕ gatePolicy: block on reproduced High fail
Merge blocked. F-2207 still reproduces on this build.GATE FAIL
08 · REVIEW AND HAND OFF

Review the Evidence. Send Engineering What They Can Act On.

AppSec reviews each validated finding in Hive before it leaves security. What reaches engineering is a ticket with the affected operation, what should have happened, what did happen, and the exact request that proved it.

01Reviewed before it is sentAppSec confirms the result, severity and owner in Hive. Nothing unproven reaches engineering.
02Operation, not just a URLThe function and service that own the weakness, so it lands with the right team.
03Expected versus observedWhat the app should have done, and what it did in the run.
04The proving requestThe exact request and response, with sensitive values redacted.
APPSEC-1192 · team-orders HighReproduced

Tenant isolation: cross-tenant read on GET /v2/orders/{id}

OPERATIONorders-api · getOrderById EXPECTED403 for any order outside the caller's tenant. OBSERVED200 with name, address and last four card digits for tenant B's orders. RUNrun_0377 · plan PL-61 · approved REVIEWEDappsec · m.okafor · sent to team-orders
GET /v2/orders/ord_8f31c2 HTTP/1.1 Authorization: Bearer <tenant-a user token> HTTP/1.1 200 OK { "tenant": "tn_b204", "customer": "[redacted]", ... }
RETEST HISTORY PR #4812 merged · replay on stagingReplay failed · fixed
09 · BUILT FOR

One Record for Security and the Teams Who Fix

Each team sees the same validation, framed for the decision it has to make.

APPSEC

Application security

A backlog ranked by proof: which findings work, which are blocked, which no route reaches.

PRODSEC

Product security

Expected boundaries per role and tenant, the scenarios run, and where the evidence stops.

ENGINEERING

Engineering and platform

The affected operation, a reproducible request, and the retest result on their own pull request.

ARCHITECTURE

Security architects

Trust boundaries, identities and dependencies, with the evidence behind each relationship.

10 · FAQ

Application Security FAQ: What AppSec Teams Ask

It tests them in the running application as your real roles and tenants, inside an approved plan. A weakness is marked reproduced only when the test worked, and every result keeps its evidence.

Book a Demo: See Which Weaknesses Are Exploitable

Bring one application and its roles. We will walk through validating authenticated behavior, business logic, tenant isolation and AI features, how your WAF and SIEM telemetry is read on each step, and how fixes are verified.

✓Validate as your real roles and tenants ✓See defenses checked on each step ✓Verify fixes before they ship
Book a Demo