softstack completed an independent penetration test of Edel Finance's perpetual futures backend, covering trading APIs, account security, AI-agent integrations, and infrastructure. Both High-severity findings were resolved and verified during the final re-test.

Client: Edel Finance

Project: Perpetual Futures Trading Backend

Industry: DeFi / Tokenized Equities / Financial Infrastructure

Service: Grey-Box Penetration Testing & Infrastructure Security Assessment

Testing Period: October 8–13, 2026

Re-Test: October 17, 2026

About Edel Finance

Edel Finance is building the infrastructure layer for programmable financial markets, with a particular focus on tokenized equities.

As traditional financial assets increasingly move onto blockchain networks, the infrastructure surrounding them must evolve.

Tokenized equities introduce opportunities beyond digital ownership. They enable programmable liquidity, composable financial products, and the integration of traditional market exposure with decentralized financial infrastructure.

Edel Finance addresses this opportunity through an ecosystem of interconnected products:

  • EdelEQUITIES provides access to tokenized equity trading and custody, supported by settlement infrastructure on the Canton Network.
  • EdelMARKETS focuses on execution, settlement, market-making, and perpetual futures infrastructure.
  • EdelRUNWAY provides capital formation and financing infrastructure for tokenized issuers.
  • EdelVAULTS is designed to support structured exposure and yield strategies built around tokenized financial assets.

Together, these components form a programmable infrastructure stack connecting assets, liquidity, exposure, and settlement.

For this engagement, softstack focused specifically on the development environment supporting Edel Finance's perpetual futures backend.

Why Security Matters for Perpetual Futures Infrastructure

Perpetual futures platforms introduce security requirements that extend beyond conventional web applications.

Their infrastructure must continuously handle financial transactions, market data, margin calculations, order execution, position management, and withdrawals.

A weakness in authentication, business logic, or infrastructure configuration can create risks even when the underlying blockchain settlement mechanisms operate correctly.

Modern trading platforms also expose an increasingly complex set of interfaces.

Alongside traditional REST APIs, trading applications may support real-time WebSocket connections, automated trading bots, and AI-powered agents.

Each interface introduces additional security considerations.

For Edel Finance, the assessment therefore needed to examine the relationship between trading operations, account authorization, API security, infrastructure, and automated integrations.

Penetration Test Scope and Methodology

softstack conducted an owner-authorized grey-box penetration test with access to the backend source code and a seeded development trading account.

This approach allowed our security specialists to combine real-world attack simulations with source-assisted verification of the underlying application behavior.

The assessment followed established methodologies, including the OWASP Web Security Testing Guide, OWASP API Security Top 10, and Penetration Testing Execution Standard (PTES).

Authentication and Authorization

We tested authentication mechanisms, bearer token validation, account isolation, and access control across trading-related operations.

Particular attention was given to object-level authorization and the prevention of unauthorized access to another user's account or financial information.

Trading Logic and Financial Operations

The review examined the core trading workflows, including:

  • Deposits and account balances
  • Order creation and execution
  • Position management and closure
  • Leverage and risk controls
  • Withdrawal operations
  • Transaction validation and replay-related protections

These checks are especially important for financial platforms where incorrect transaction handling or inconsistent authorization can affect account balances and trading activity.

API and Infrastructure Security

Our specialists reviewed the platform's REST APIs, WebSocket integrations, administrative interfaces, network exposure, and application security configurations.

The assessment also covered infrastructure-as-code configurations, container security, dependency management, and CI/CD workflows.

AI-Agent and MCP Security

An important part of the engagement involved Edel Finance's integration of the Model Context Protocol (MCP).

MCP enables compatible AI agents to interact with supported application functionality through standardized tools.

For trading applications, this creates a security boundary between AI-generated instructions, user-controlled data, account permissions, and actions that may affect financial positions.

Our assessment examined MCP authentication, account isolation, response handling, and the handling of potentially malicious instructions embedded in external data.

Security Assessment Results

The penetration test identified 27 findings across the assessed infrastructure.

  • Critical: 0
  • High: 2
  • Medium: 5
  • Low: 17
  • Informational: 3
  • Total: 27

The findings covered several areas of application and infrastructure security.

Software Dependencies and Supply-Chain Security

One High-severity finding involved backend dependencies using software versions affected by known security vulnerabilities.

Maintaining secure dependency versions is an important part of protecting modern financial infrastructure, particularly when applications depend on multiple external libraries and services.

AI-Agent Security and Prompt Injection

The second High-severity finding involved a potential cross-user prompt-injection scenario affecting the MCP trade-history interface.

The finding highlighted the importance of maintaining strict boundaries between untrusted application data and instructions processed by AI agents.

As AI-assisted trading applications become more common, this represents an increasingly relevant area of financial cybersecurity.

Infrastructure and API Hardening

Additional findings involved network exposure, application-layer security controls, API input validation, and infrastructure configuration.

The review also identified opportunities to strengthen security headers, container configurations, and automated development workflows.

These issues were documented with remediation recommendations and addressed during the engagement.

Positive Security Controls

Identifying vulnerabilities was only one part of the assessment.

Our testing also confirmed several important security controls were functioning correctly in the tested environment.

  • Account isolation and authorization: Object-level authorization held across the tested account-scoped operations, including batch order functionality.
  • Authentication: Bearer token validation behaved as intended during the assessment.
  • Financial transaction integrity: The tested financial workflow, from deposits through order placement, execution, and position closure, reconciled correctly.
  • Fail-closed behavior: Commands capable of moving value rejected operations when required upstream dependencies were unavailable.

For a financial trading platform, the correct implementation of these controls is particularly important.

Remediation and Re-Testing: 25 Findings Fixed

Following the initial assessment, the Edel Finance engineering team worked closely with softstack to address the identified findings.

Between October 11 and 18, the team merged 23 remediation pull requests into the backend.

A dedicated re-test was performed on October 17, 2026, examining the implemented fixes and relevant infrastructure configurations.

Final Results

  • 25 findings successfully remediated and verified
  • 2 Low-severity findings formally accepted as residual risk
  • 0 findings remaining open

Both High-severity vulnerabilities and all five Medium-severity findings were successfully resolved.

The two accepted residual risks concerned the intended scope and behavior of Cloudflare managed-challenge controls.

This distinction matters: a finding can be fixed, or the associated risk can be formally accepted following review. Both outcomes should be clearly documented rather than presented as identical.

The final report provides a traceable record of the original findings, the remediation work, and the subsequent verification.

Full penetration test report: softstack Edel Finance penetration test report

What This Engagement Demonstrates

The security requirements of financial infrastructure are changing.

As tokenized assets become integrated with automated trading, blockchain settlement, and AI-powered applications, security assessments must cover the complete technical interaction between these components.

For Edel Finance, this meant going beyond conventional web application testing.

The engagement examined trading workflows, API authorization, AI-agent interfaces, infrastructure exposure, and software supply-chain security within the defined development environment.

The collaboration between softstack and the Edel Finance engineering team resulted in 25 verified fixes, including both High-severity findings, and no remaining open findings.

The assessment did not cover every Edel Finance product or its complete production infrastructure. Future changes and deployments require their own security controls and ongoing review.

Security is not a one-time checkbox. It is an engineering process built on identifying risks, implementing improvements, and independently verifying the results.

About softstack

softstack is a Germany-based cybersecurity and software development company specializing in blockchain technology and digital asset infrastructure.

Founded in 2017, softstack provides smart contract security audits, penetration testing, digital asset risk assessments, and specialized security consulting for Web3 protocols, financial institutions, and digital asset businesses.

Our security assessments combine manual expertise, automated testing, and practical remediation support to help organizations identify and reduce technical risks.

Building trading infrastructure, tokenized asset platforms, or AI-powered financial applications?

Contact our security team at hello@softstack.io or visit softstack.io to discuss an independent security assessment.