Identity and access
Authentication, recovery, sessions, tokens, roles, object access, tenant separation, ownership, and entitlements.
Security Assessments / Web Application
Test where trust can break.
Authenticated, source-assisted gray-box testing across one application, its first-party interfaces, identities, tenant boundaries, data, integrations, and critical workflows.
Where can an attacker cross a boundary the system was meant to preserve?
The assessment follows identities, authority, data, and state across the application—not just isolated endpoints. Runtime behavior and targeted source analysis are combined to expose failures that exist between components and control paths.
Application and APIs
Input handling, injection, unsafe execution, files, outbound requests, browser security, APIs, and unintended exposure.
Critical workflows
State transitions, replay, concurrency, limits, approvals, impersonation, overrides, integrations, and background paths.
Source-assisted controls
Targeted tracing of security-critical implementation paths to guide testing, find alternates, and connect behavior to root cause.
Coverage is risk-based and representative. Source access increases depth, but the engagement is not a full code audit and does not promise exhaustive endpoint, role, workflow, dependency, configuration, or source-line coverage.
Model the system. Then challenge it.
The work starts with the application's actual trust boundaries and security-critical behavior. Testing follows the paths where failure would matter most.
- 01Map trust
Build a working model of interfaces, roles, tenants, sensitive operations, data, administrative paths, integrations, and asynchronous processing.
- 02Exercise authority
Test authentication and authorization independently of browser restrictions across representative objects, actions, roles, tenants, and interfaces.
- 03Break workflows
Challenge state, sequence, replay, concurrency, approvals, limits, privileged functions, callbacks, queues, and alternate execution paths.
- 04Trace controls
Use source to guide runtime testing and inspect selected security-critical implementation paths—not to inflate the engagement into a line-by-line audit.
- 05Validate and synthesize
Close candidates with bounded evidence, connect compound paths, and prioritize remediation by expected risk reduction.
Findings with context and proof.
The report makes the affected boundary, evidence, consequence, uncertainty, and remediation direction explicit.
Critical notification
Validated Critical findings are communicated immediately rather than held for the final report.
Written report
Executive conclusion, tested coverage, limitations, evidence-backed findings, severity rationale, material attack paths, and prioritized remediation.
Engineering readout
A focused discussion of material findings, attack paths, and remediation priorities when live discussion adds value.
Define the scope.
Each engagement receives a fixed scope and fixed quote before work begins. Breadth and complexity shape the quote; evidence integrity does not.
To scope the workAccess + context
Provide representative accounts and tenant contexts, approved test data, application access, the relevant source revision, available API definitions, and a technical contact.
One coherent application includes its first-party web, API, administrative, and supporting interfaces within the agreed boundary.
Discuss an assessment
What do you need assessed?
A brief outline is enough to start. We’ll follow up by email to discuss fit, scope, and timing.
Scope and a fixed quote are agreed before work begins.
Prefer email? mark@inferencesecurity.ai