Named contributor to the Upanzi Network’s published Africa report on the security of public e-government infrastructure, a large-scale automated assessment of 21,782 discoverable subdomains across all 54 African states for misconfigurations, outdated components, cryptographic weaknesses, and data exposure. I led the HTTP-methods enumeration work, flagging dangerous exposed methods, and built the custom Python automation that powered mass asset discovery and scanning.
Emmanuel Tega
Agbragu
I work both sides of security. I break web apps, APIs, and AI-backed services as a penetration tester, and I build the defensive infrastructure (SIEM, Wazuh, IDS/IPS, EDR, DLP) that keeps regulated environments defensible. Either way, I turn findings into clear, actionable risk-based remediation that engineers and executives can both act on.
Security with a paper trail
I’m a cybersecurity engineer with 4+ years across offensive and defensive security. On the offensive side I focus on web and API penetration testing, AI/LLM-backed services as well as Infrastructure security. On the defensive side I’ve deployed and tuned the infrastructure that protects regulated environments: IDS/IPS, EDR, firewalls, privileged access management, and data loss prevention, alongside SOC and detection tools.
My most recent role was security researcher at the Upanzi Network (CMU–Africa), where I conducted penetration test on digital public infrastructure: APIs, chatbots, e-commerce platforms, and data portals. I worked the full loop, from recon and enumeration through exploitation, proof of concept, and remediation, and I wrote findings so that an engineer can fix them and an executive can understand the risk. Before that, I built security infrastructure for banking clients in Lagos.
I run against OWASP Top 10, the OWASP API and LLM Top 10s, and NIST risk methodology, and I lean heavily on Python and Bash to automate at scale.
Selected Projects
A sample of web and API penetration tests I worked on with the Upanzi Network security team at CMU–Africa. Targets, hostnames, and exploit specifics are intentionally withheld; these summaries describe vulnerability classes, impact, and remediation only.
Open Data Portal
Enumerated APIs and client-side JavaScript, mapped hidden endpoints, and tested an AI description-generation feature. Found an Indirect Prompt Injection (LLM01) flaw where instructions embedded in uploaded PDFs hijacked the model, plus unauthenticated PII exposure and a missing-input-validation DoS in the datasets API.
MojaShop
Authentication, authorization, and business-logic testing of a shopping platform. Found a Mass Assignment flaw in registration that let any user self-provision an admin account, then chained it to Broken Object Level Authorization (BOLA/IDOR) giving full create/update/delete control over products and exposure of customer and transaction PII.
Policy Analyser
Discovered an unauthenticated backend API exposed on a non-standard port with public API docs. Demonstrated full database retrieval and successful delete operations against policy data, plus missing security headers enabling clickjacking and XSS. Findings were remediated: authentication and authorization were added to the API endpoints.
NiD Chatbot Backend
Tested an LLM chatbot backend and its Telegram bot codebase, combining live API testing with static code review. Found unauthenticated endpoints allowing cache abuse (DoS), weak cryptographic primitives (MD5/SHA1), an outdated Flask dependency with a known CVE, and schema disclosure via public API docs. Prompt-injection attempts were correctly rejected.
// Note on disclosure. All engagements above were authorized assessments. Live targets, internal IP addresses, credentials, and step-by-step exploitation details are deliberately omitted from this public page. Full sanitized reports and proof-of-concept material are available to prospective employers on request.
Experience
Four-plus years spanning offensive research, teaching, and the defensive security infrastructure that protects regulated banking environments.
Jan 2025 – Apr 2026 · Offensive
Sep 2024 – Jan 2025 · Offensive + Defensive
Apr 2021 – Jun 2023 · Defensive
What I work with
Offensive security & VAPT
LLM / AI security
Vuln management & recon
Defensive & blue team
Scripting & automation
Frameworks & standards
Certifications & achievements
// Achievements. Top 10 finalist, Greenfist CTF 2023 · Mastercard Foundation Scholar 2023.
Education & coursework
Systems programming
Let’s talk security
Open to Security Engineer, Penetration Testing, Red Team, AppSec, and SOC analyst related roles, including relocation. Reach out and I’ll share full sanitized reports and references.