Skip to main content
Security Testing

Web Application Penetration Testing

Manual, researcher-led web application pentesting across your full attack surface, with every finding verified by a person

Every finding human-verified Scoped to your goals Results tracked in Talon
1-2 weeksTypical testing window
Custom scopeQuoted after a short scoping call
24 to 48 hoursRetest turnaround on verified fixes
Human sign-offLory finds, a pentester signs off
Overview

What this engagement covers

The service

Modern web applications are complex distributed systems combining client-side single page frameworks (React, Vue, Next.js), asynchronous APIs (REST, GraphQL, WebSocket), microservices, and multi-tenant cloud datastores. Automated DAST scanners only detect superficial signature-based flaws and flood security backlogs with false positives while missing over 70% of critical vulnerabilities—such as tenant-crossing Broken Object Level Authorization (BOLA), multi-step business logic bypasses, and race conditions. Lorikeet Security's web application penetration testing combines proprietary automated attack surface footprinting with rigorous, adversarial manual exploitation by senior offensive researchers. Every finding is verified with working proof-of-concept exploits, CVSS v3.1/v4 scoring, and automated PR remediation via Talon MCP, all backed by zero false positives and continuous validation under Talon PTaaS.

What we test

We thoroughly assess modern multi-tenant SaaS applications, single-page architectures (Next.js, Remix, React, Vue, SvelteKit), serverless backends, GraphQL federations, and REST API layers. Testing covers authentication mechanics (OAuth 2.0, OpenID Connect, SAML 2.0, WebAuthn, JWT algorithm validation), session lifecycle (fixation, entropy, revocation), multi-tenant authorization boundaries (IDOR, BOLA, horizontal/vertical privilege escalation), business logic workflows (checkout, subscription downgrades, state-machine manipulation, financial rounding flaws), input processing (SQLi, NoSQLi, blind SSRF, SSTI, command injection, path traversal), client-side security (DOM XSS, prototype pollution, CORS misconfigurations, CSP bypasses), file processing pipelines (unrestricted uploads, ImageMagick/PDF rendering RCE), and WebSocket communication channels. Testing aligns with OWASP Top 10 (2021), OWASP ASVS v4.0.3 Level 2/3, and the OWASP Web Security Testing Guide (WSTG v4.2).

Attack surface

Where we focus

The attack paths we spend the most time on, because they are where real breaches start.

Authorization & Multi-Tenancy

IDOR / BOLA - CWE-639 / CWE-284

Manipulating resource identifiers, tenant IDs, and routing parameters in stateful API requests to breach organizational tenancy isolation and exfiltrate cross-customer data.

Authentication Bypass & JWT Algorithm Confusion

CWE-287 / CWE-347

Exploiting token validation flaws including alg:none stripping, RS256-to-HS256 public key HMAC signing confusion, missing expiration checks, and race conditions in passwordless/OAuth login flows.

Server-Side Request Forgery (SSRF) to Cloud Metadata

CWE-918

Abusing unvalidated webhooks, document/PDF generation services, or image fetchers to query link-local addresses (169.254.169.254) and harvest temporary AWS/GCP IAM STS credentials.

Business Logic & Concurrency / Race Conditions

CWE-840 / CWE-362

Leveraging HTTP/2 single-packet multiplexing and asynchronous state lag to exploit Time-of-Check to Time-of-Use (TOCTOU) flaws, enabling double-spending, coupon stacking, and authorization bypasses.

Server-Side Template Injection (SSTI) & Deserialization RCE

CWE-1336 / CWE-502

Injecting malicious template expressions into Jinja2, Thymeleaf, or Freemarker engines, or weaponizing unsafe object deserialization in Node.js, Python, or Java to execute shell commands.

Client-Side State Poisoning & CORS / CSP Bypass

CWE-942 / CWE-79 / CWE-1321

Targeting permissive Access-Control-Allow-Origin reflections with credentials, prototype pollution via URL query parameters, and strict CSP script-src bypasses to hijack authenticated browser sessions.

Testing depth

Choose how much we know going in

Adversary Simulation

Black Box Testing

Zero-knowledge testing simulating an unauthenticated external adversary targeting your public attack surface, unlinked routes, forgotten subdomains, and authentication portals.

Maximum Depth

White Box / Hybrid Testing

Testers are provided with architecture blueprints, API schemas, and source code access alongside live staging environments. Delivers exhaustive vulnerability discovery for pre-launch architectures and high-risk applications.

Methodology

How we run it

Aligned with the OWASP Web Security Testing Guide (WSTG v4.2) and the Penetration Testing Execution Standard (PTES), our testing operates under an assumed-adversary mindset. We begin with automated surface footprinting and parameter discovery, followed by hands-on manual abuse of application business logic and authorization boundaries across multiple role tiers. Senior offensive researchers manually interrogate every endpoint to chain low-severity anomalies into high-impact remote exploits without causing production downtime. Lorikeet Security maintains a strict zero false positive philosophy: no unverified scanner noise is ever reported. Every vulnerability is independently verified with reproducible curl commands and raw HTTP evidence. All findings stream in real-time to your Talon PTaaS portal with bi-directional Jira/GitHub ticket synchronization and automated PR code fixes via Talon MCP, concluding with a formal human-signed attestation letter and complimentary retesting.

  1. 01

    Reconnaissance & Attack Surface Mapping

    We map the full digital perimeter of your web application, discovering hidden parameters, subdomains, forgotten endpoints, legacy routes, API documentation, and third-party integrations using active probing and OSINT.

  2. 02

    Automated Baseline Scanning & Discovery

    We deploy proprietary scanners and industry-standard tools to establish an architectural baseline, identify known CVEs, outdated libraries, and misconfigurations while eliminating scanning noise through manual validation.

  3. 03

    Authentication & Session Security Review

    In-depth manual testing of login logic, password reset mechanisms, Multi-Factor Authentication (MFA) bypasses, JWT signature verification, session token randomness, timeout handling, and session fixation vulnerabilities.

  4. 04

    Authorization & Access Control Testing (IDOR / BOLA)

    Manual testing across multi-tenant boundaries and privilege tiers. We probe for Insecure Direct Object References (IDOR), horizontal privilege escalation (user-to-user), and vertical privilege escalation (user-to-admin).

  5. 05

    Business Logic Flaws & Workflow Abuse

    We deconstruct multi-step business workflows (e.g. checkout, payment gateway callbacks, coupon application, voting, onboarding) to find state manipulation flaws that automated vulnerability scanners are blind to.

  6. 06

    Input Validation, Injection & Deserialization

    Manual injection testing targeting SQL Injection, NoSQL Injection, Server-Side Template Injection (SSTI), Command Injection, Cross-Site Scripting (XSS), XML External Entities (XXE), and insecure object deserialization.

  7. 07

    API, Microservices & GraphQL Assessment

    We inspect REST and GraphQL endpoints consumed by your web client, testing for Broken Object Level Authorization (BOLA), mass assignment, batching attacks, rate-limit bypasses, and unauthenticated schema introspection.

  8. 08

    Exploitation, Evidence Capture & Chained Impact

    Every valid exposure is manually weaponized in a non-destructive manner to demonstrate real-world business impact. We capture reproduction curl commands, HTTP request/response logs, and video PoCs for your developers.

Deliverables

What you get, and what we usually find

Findings land in Talon as they are confirmed, not only in a PDF at the end, and retests are part of the engagement.

You receive

  • Executive Summary with Business Impact Analysis & Board-Level Risk Posture Scorecard
  • Technical Vulnerability Report with Verified CVSS v3.1 and CVSS v4.0 Metric Scoring
  • Real-Time Talon Interactive Dossier with Live Vulnerability Tracking & Asset Inventory
  • Copy-Paste curl Reproduction Commands and Raw HTTP Request/Response Evidence Payloads
  • Automated Remediation Pull Requests (PRs) via Talon MCP Server for Instant Developer Merging
  • Line-by-Line Developer Remediation Guidance with Framework-Specific Code Patches (React, Node, Python, Go)
  • Auditor-Ready Compliance Crosswalk (SOC 2 CC7.1/CC7.2, PCI-DSS 4.0 Req 11.4, HIPAA, ISO 27001 A.8.8)
  • Formal, Human-Signed Letter of Attestation for Enterprise Vendor Security Reviews & Customers
  • Post-Remediation Verification Retest Report validating Resolved Vulnerabilities with Clean Attestation

Common findings

The issues this engagement surfaces most often.

Broken Object Level Authorization (BOLA / IDOR)Authentication Bypass & Session FixationBusiness Logic & Multi-Step Workflow AbuseSQL Injection & NoSQL Database InjectionServer-Side Request Forgery (SSRF to Cloud Metadata)Cross-Site Scripting (Stored & DOM XSS)Server-Side Template Injection (SSTI)Mass Assignment & GraphQL Introspection Flaws
Compliance

Evidence your auditor will accept

Findings are mapped to the controls below, so the report drops straight into your audit package.

FrameworkControlWhat the pentest evidences
PCI DSS v4.0 Requirement 11.4.3 Annual external and internal penetration testing of the cardholder data environment (CDE) and web-facing application layer by qualified personnel.
SOC 2 Type II CC4.1, CC4.2, CC7.1 Formal penetration testing evidence and remediation tracking formatted specifically for CPA auditor workpapers to validate continuous security controls.
HIPAA Security Rule 45 CFR § 164.308(a)(1) Technical evaluation and vulnerability assessment verifying that electronic Protected Health Information (ePHI) data stores and workflows cannot be compromised.
ISO/IEC 27001:2022 Control A.8.8 Technical vulnerability management through independent penetration testing, risk-prioritized findings, and verified corrective action plans.
FedRAMP / NIST NIST SP 800-53 Rev 5 CA-8 Adversary-driven penetration testing simulating realistic threat vectors, satisfying federal agency authorization and continuous monitoring mandates.

Who this is for

Series A to Growth-Stage B2B & B2C SaaS Platforms
Fintech, Payment Gateways & Banking Applications (PCI-DSS 4.0)
Digital Health, Telemedicine & HealthTech Portals (HIPAA)
Enterprise Customer-Facing Portals & Multi-Tenant Marketplaces
Pre-IPO Cloud Software Companies Preparing for Vendor Due Diligence
High-Volume E-Commerce & Transaction-Heavy Applications
Startups Completing Annual SOC 2 Type II or ISO 27001 Audits
OWASP Top 10PCI-DSSHIPAASOC 2ISO 27001
Pricing

Pentesting included in Talon

This engagement is scoped and quoted separately after a short call. Your results live in Talon, and these plans cover your ongoing testing.

Essentials

Continuous autonomous testing for fast-shipping teams.

$1,999/yr
$166.58/mo billed annually, or $199 monthly
  • 250 Lory credits / month
  • Autonomous web, API & network testing
  • Real-time findings with reproduction steps
  • Unlimited 1-click retests
  • Talon MCP server for Cursor & Claude Code
  • One-click Jira & GitHub tickets
Start Essentials

Enterprise

Full-estate coverage with priority SLAs and advisory.

$9,999/yr
$833.25/mo billed annually, or $999 monthly
  • 1x annual pentest, up to 5 assets
  • Any asset type: web, mobile, cloud, AD, AI, IoT
  • 1,500 Lory credits / month
  • Lory Mesh private network connector
  • Priority 24-hour retest SLA
  • Multi-framework attestation pack
Choose Enterprise
Talon Free Workspace, $0

View past reports and findings, connect your IDE through the Talon MCP server, and run Lory on pay-as-you-go credits.

Why Lorikeet Security

Talon vs. traditional firms vs. scanners

 Lorikeet Security + TalonTraditional firmsAutomated scanners
How often you are testedLory tests continuously; our team runs your annual pentestOnce a yearScheduled scans
Findings verified by a humanEvery one, countersignedYes, for the tested scopeNo
When you see findingsLive in Talon as they are confirmedIn a PDF weeks after testingInstantly, unverified
Retesting fixesUnlimited, 1-click, 24 to 48 hoursPaid change orderRe-scan only
Fix workflowJira, GitHub & your IDE via MCPCopy from the PDF by handGeneric advice
Accepted by auditorsYes, with an attestation letterYesRarely on its own
FAQ

Common questions

Automated DAST scanners match known signatures and response patterns, so they routinely miss authorization bypasses (BOLA/IDOR), multi-step business logic flaws, race conditions, and chained exploits. They also produce noise your engineers have to triage. Our pentesters test manually on top of automated discovery, and every finding is checked and signed off by a Lorikeet Security pentester before you see it.

Yes. Before testing we agree Rules of Engagement, set up test accounts at each privilege level, and tag our traffic with a unique header (X-Lorikeet-Pentest: true) so your SIEM or WAF can filter it. We avoid destructive payloads such as denial of service, and risky flows like live billing or bulk email are coordinated with your engineering leads.

We ask for at least two tenant organizations, each with several roles (for example owner, member, and admin). Our pentesters capture API traffic and swap tokens, cookies, and object IDs across tenant and role boundaries. This exposes horizontal access (Org A reading Org B data) and vertical escalation (a member performing admin-only actions such as billing or user management).

You get a signed Letter of Attestation stating scope, methodology (aligned with NIST SP 800-115 and OWASP WSTG), testing dates, and summary results. The technical report maps findings to SOC 2 Trust Services Criteria (CC4.1, CC7.1, CC7.2) and PCI DSS v4.0 Requirement 11.4. Once fixes are retested, the report shows the verified status for your auditor.

Every finding comes with written fix guidance for your stack, such as Next.js, Express, Django, or FastAPI. The Talon MCP server lets your engineers read findings and request retests from Cursor or Claude Code. Your team writes and ships the fix, and a pentester verifies it when you request the retest.

Yes. Retests are included and unlimited on Talon plans. When a fix is deployed, request a retest in one click from the finding in Talon, and a pentester re-runs the original exploit to confirm it is closed, typically within 24 to 48 hours (24 hours priority on Enterprise). You can also create a Jira, GitHub, GitLab, or Azure DevOps ticket from any finding in one click.

We can typically begin testing within 2 to 3 business days of scope sign-off, and active testing usually takes 1 to 2 weeks depending on application size and API surface. Findings are published to your Talon portal as they are verified, so your team can start fixing critical issues before the final report is delivered.

Lory waving

Hi, I'm Lory! Need help finding the right service? Click to chat!