Bug bounty hunter · Security researcher · Automation engineer
I build automated systems that perform continuous, authorized security research — multi-model AI pipelines, target inventory, recon, lead discovery, triage, and verification — orchestrated with GitHub Actions.
Software that turns bug-bounty hunting into a repeatable, measurable engineering discipline instead of a series of manual clicks:
- Autonomous hunting pipelines — scheduled, multi-model agents that move through recon → surface → hypothesis → triage → verification, one phase per cycle, with state and a shared knowledge base.
- Multi-model analysis — prompt-driven analyst, triager, and verifier roles; outputs are ranked by confidence and gated through a 7-Question validation gate before anything is reported.
- Passive-first, scope-bound testing — every pipeline is bound to an explicit authorized scope, uses GET/HEAD/OPTIONS-only probes where rules require it, and enforces rate limits, politeness budgets, and 403/429 grace-stops.
- Target inventory & verification — per-target asset lists, live-host checks, and an honest distinction between leads (candidate findings) and validated bugs.
The *-hunt repositories are concrete deployments of this model against specific authorized bug-bounty programs.
| Project | Description |
|---|---|
| DedupeAI | Burp Suite (Montoya) extension: deduplicates HTTP history into an AI-ready unique-request feed, color-codes attacker/victim traffic by listener port, and ships the set to Claude Code — built for multi-account IDOR/BOLA testing. |
| oniontui | Terminal AI assistant that browses and searches the web — including .onion hidden services — through Tor, with plan-first agents, working memory, and circuit rotation. |
| Js-Scanner | AI-powered JS security audit tool: crawls a site's scripts, extracts endpoints/secrets/JWTs, and produces an interactive report. |
| gladia-hunt | 24/7 multi-model bug-hunting automation bound to the Gladia authorized scope. |
| threema-hunt | 24/7 passive, read-only multi-model hunting pipeline for the Threema program. |
Web2 · IDOR · broken access control · auth/ATO chains · business logic · SSRF · injection
Mobile · Android/iOS app assessment (jadx, Frida, objection, MobSF)
AI/LLM · prompt injection · RAG/vector-store poisoning · agentic AI security (ASI01–ASI10)
Web3 · smart-contract auditing (Foundry, Slither, Echidna)
Cloud / infra · AWS/GCP/K8s misconfiguration · post-credential privilege analysis
GitHub Actions · opencode multi-model agents · Python · recon (subfinder, dnsx, httpx) · Burp Suite / Montoya · Frida · jadx · Foundry · Slither · orchestrating scheduled, stateful, verifiable research.
- Scope first — every pipeline is bound to an explicit authorized target with exclusions and safe defaults.
- Passive before active — read-only probes where program rules require it.
- Leads ≠ findings — candidate hypotheses are triaged by a second model and validated before reporting.
- Humans decide — models propose; I verify and submit.
Hack legally. Report responsibly. 🔒
| Metric | Value |
|---|---|
| Followers | 5 |
| Public repos | 70 |
| Total stars | 8 |
Recently updated:
- roobet-hunt — 24/7 deep bug-hunting automation for Roobet (bugs.olivermaicher.eu)
- hornbach-hunt — 24/7 deep bug-hunting automation for HORNBACH Holding AG & Co. KGaA / HORNBACH Group (bugs.olivermaicher.eu)
- daimler-truck-hunt — 24/7 deep bug-hunting automation for Daimler Truck Holding AG (bugs.olivermaicher.eu)
- questnet-gmbh-hunt — 24/7 deep bug-hunting automation for Questnet GmbH (bugs.olivermaicher.eu)
- obi-hunt — 24/7 deep bug-hunting automation for OBI Group Holding SE & Co. KGaA (bugs.olivermaicher.eu)
Last refreshed: 2026-09-03 08:52 UTC — auto-updated daily by GitHub Actions (
.github/workflows/profile-stats.yml)