The docket / Five service lines

    Everything we test — not just the web app.

    Most engagements are sold as "a pen test" and quietly mean one web application. Attackers do not respect that boundary. Below is the complete docket: what each line covers, what we routinely find, and what lands in your hands at the end.

    The docketExternal networkInternal networkWeb applicationAPICloud configurationExternal estate3 white box · 3 limited access6 service lines

    Service lines

    01 — 06

    01 / Perimeter

    External network penetration test

    Everything of yours that answers from the public internet — hosts, services, exposed management interfaces, and the forgotten things nobody remembers standing up.

    Access · Black-box · Gray-box

    What we test

    • Full-range host and service discovery across owned ranges and cloud-assigned addresses
    • Exposed administrative interfaces, remote access, and management planes
    • Perimeter device and VPN configuration, including authentication weaknesses
    • Mail, DNS, and certificate hygiene where it creates an exploitable path
    • Credential exposure and password spraying against internet-facing authentication
    • Verified exploitation of anything reachable, with a proof-of-concept

    Findings we routinely see

    • A staging host with production data, reachable and unauthenticated
    • A management interface exposed to the internet with default credentials
    • A VPN appliance behind on patches with a known authentication bypass

    What you receive

    • Confirmed external asset inventory — including assets you did not know you owned
    • Exploited findings with reproduction steps and evidence
    • Perimeter-hardening sequence in remediation order

    02 / Post-access

    Internal network penetration test

    What an attacker does after the first foothold — the part most organizations have never actually measured. We start inside, from a standard workstation or a network drop, and see how far the network lets us go.

    Access · Gray-box · White-box

    What we test

    • Lateral movement from a standard user position toward high-value systems
    • Active Directory and identity attack paths — delegation, ACLs, credential reuse
    • Privilege escalation on hosts and across trust boundaries
    • Network segmentation validation: can the user VLAN reach production or backups
    • Credential harvesting from memory, shares, scripts, and service accounts
    • Detection checkpoints — what your tooling saw, and when

    Findings we routinely see

    • Domain administrator reached in under a day from a standard user account
    • Flat network where a workstation can open a database port in production
    • Service account with a reused password and privileges nobody documented

    What you receive

    • Attack-path diagram from starting position to domain or data compromise
    • Segmentation findings mapped to the boundaries you believed existed
    • Detection gap timeline — the actions that should have alerted and did not

    03 / Application

    Web application penetration test

    The application itself, tested as each of its user types. Not a scanner pass — the logic, the boundaries between tenants and roles, and the workflows a scanner cannot reason about.

    Access · Black-box · Gray-box · White-box

    What we test

    • Authentication, session handling, password reset, and multi-factor bypass
    • Authorization across roles and tenants, including horizontal access between accounts
    • Business-logic abuse in the workflows that move money, data, or permissions
    • Injection, deserialization, template, and file-handling paths through to impact
    • Client-side issues that lead to account takeover rather than a theoretical alert
    • Administrative and support tooling, which is usually the weakest surface

    Findings we routinely see

    • One tenant reading another tenant's records by changing an identifier
    • A support tool that lets a low-privilege operator elevate to full administrator
    • A password reset flow that can be completed without controlling the mailbox

    What you receive

    • Findings ranked by exploited impact, not by scanner severity
    • Reproduction steps a developer can follow without asking a question
    • Retest of remediated findings with an updated, shareable report

    04 / Interface

    API penetration test

    REST, GraphQL, and internal service interfaces tested directly, without the front end in the way. APIs enforce authorization far less consistently than the UI that calls them.

    Access · Gray-box · White-box

    What we test

    • Object-level and function-level authorization on every exposed operation
    • Authentication and token handling, including scope, expiry, and replay
    • GraphQL introspection, query depth, batching, and resolver-level access control
    • Mass assignment, parameter tampering, and unintended field exposure
    • Rate limiting and abuse paths against expensive or sensitive operations
    • Undocumented and deprecated endpoints still answering in production

    Findings we routinely see

    • An endpoint returning every customer record when the identifier is incremented
    • A deprecated v1 route still live, and still missing the authorization added in v2
    • A token scoped for read that a write operation accepts without checking

    What you receive

    • Endpoint-level coverage record — what was tested and what each operation allowed
    • Authorization matrix showing which roles could reach which operations
    • Machine-readable finding file for your issue tracker

    05 / Cloud

    Cloud configuration review

    AWS, Azure, and GCP accounts reviewed the way an attacker reads them: as a set of identities, permissions, and reachable resources rather than as a compliance checklist.

    Access · Gray-box · White-box

    What we test

    • Identity and permission analysis — over-privileged roles, escalation paths, stale principals
    • Storage exposure across buckets, blobs, snapshots, and backups
    • Network posture: security groups, peering, and unintended reachability
    • Key, secret, and credential handling across services and pipelines
    • Logging and monitoring coverage — whether an attacker's actions would leave a record
    • Workload configuration for containers, serverless functions, and managed databases

    Findings we routinely see

    • A build role that can assume a production role nobody intended it to reach
    • A database snapshot shared publicly during a migration and never unshared
    • Audit logging disabled in the one account that holds customer data

    What you receive

    • Privilege-escalation paths drawn from the account's actual policy graph
    • Exposure inventory ordered by what is reachable and what it holds
    • Remediation sequence written against your infrastructure-as-code, not the console

    06 / Estate

    External estate assessment

    Every internet-facing application you own, not the two you can afford to scope by hand. Autonomous agents test the whole estate in parallel from where real attackers start — no credentials, no source code — and our engineers confirm every finding before it reaches you.

    Access · Black-box

    What we test

    • Discovery of your external application estate, approved by you as a target list before testing starts
    • Parallel uncredentialed testing across tens to hundreds of applications and APIs
    • Forgotten staging hosts, abandoned subdomains, and shadow IT nobody inventoried
    • Testing with your web application firewall in place, including bypass attempts
    • Pivot paths from low-priority applications toward the ones that matter
    • Human verification of every exploit before it is reported

    Findings we routinely see

    • A marketing microsite on an old framework with a remote code execution path
    • An acquired company's login portal still live and still sharing an identity provider
    • A forgotten API host answering without authentication behind a misconfigured firewall rule

    What you receive

    • Confirmed external application inventory, scoped and approved by you
    • Exploitable findings only, each with a working proof-of-concept
    • Shortlist of applications that warrant a deep, credentialed test next
    Read more: Autonomous estate assessment, human-verified →

    Engagement depth

    Scope, not price
    Scope estimatorThree answers · one recommendation

    Surfaces in scope

    Environment

    Retest

    Recommended tierFocusedFocused engagements · 5 to 7 business days
    Scope this →
    01 / Depth

    Focused

    One surface

    A single service line, tested end to end in a short window. The right shape for a first test, an annual re-check, or a specific auditor or insurer request.

    02 / Depth

    Standard

    Two to three related surfaces

    Adjacent surfaces tested together so findings can chain — an app with the API behind it, or an external perimeter with the cloud account it fronts.

    03 / Depth

    Extended

    Multi-surface, chained across environments

    The full path an adversary would actually take: perimeter to internal, application to identity, cloud to workload. Findings are reported as chains, not as a list.

    Depth is a scoping decision, not a package. Every quote is fixed-scope and confirmed after a 30-minute call — scope your engagement.

    Access models

    How much we start with
    BLACK

    Black-box

    No credentials, no diagrams. We start where an unauthenticated attacker starts and work inward from public information alone.

    GRAY

    Gray-box

    Standard user credentials plus architecture context. The most common choice: it removes the guessing and spends the window on depth instead of discovery.

    WHITE

    White-box

    Full credentials, configuration, and source or infrastructure-as-code where relevant. Maximum coverage per day of effort.

    Coverage matrix

    Line × depth
    BLACKGRAYWHITEExternal networkInternal networkWeb applicationAPICloud configurationExternal estate
    01 · External networkBLACK · GRAY
    FocusedPerimeter discovery and exploitation of internet-facing services.
    StandardPerimeter plus the cloud account or application estate behind it.
    ExtendedPerimeter used as the entry point for a full chained attack path.
    02 · Internal networkGRAY · WHITE
    FocusedSingle site or segment, assumed-breach start from one position.
    StandardInternal estate plus identity infrastructure and segmentation validation.
    ExtendedMulti-site internal testing chained with perimeter or cloud entry.
    03 · Web applicationBLACK · GRAY · WHITE
    FocusedOne application, its primary roles, and its authenticated surface.
    StandardApplication plus its API and the cloud configuration behind it.
    ExtendedApplication estate tested as a set, with cross-application chaining.
    04 · APIGRAY · WHITE
    FocusedOne API surface, its documented operations, and its auth model.
    StandardAPI plus the application consuming it and the identity provider behind it.
    ExtendedFull service mesh or multi-API estate, including internal service-to-service calls.
    05 · Cloud configurationGRAY · WHITE
    FocusedOne account or subscription, identity and exposure review.
    StandardMulti-account organization with cross-account trust analysis.
    ExtendedCloud estate tested as an attack path from perimeter through workload to data.
    06 · External estateBLACK
    FocusedUp to a few dozen public applications, uncredentialed.
    StandardFull external estate plus verified pivot paths between applications.
    ExtendedFull estate sweep followed by deep credentialed tests on the riskiest applications.
    Access models per line · BLACK = black-box · GRAY = gray-box · WHITE = white-box

    Next step

    Tell us what exists, and we will tell you what to test.

    Scoping is a 30-minute conversation. Pick the lines you already know you need — we will flag the ones you did not think to ask about.

    Request a pen test scope call →