Security engineering practice notes from Sein Linn (wh1sky02), Singapore. The work sits between offensive security research and applied AI, and a running habit shapes everything here: when a job needs a tool, build the tool, publish it.
The portfolio at 1m1ssher.github.io reads like the full paper; these are the entries worth knowing first.
| Entry | Why it matters |
|---|---|
| copy-fail-python | Python proof-of-concept for CVE-2026-31431, a Linux kernel privilege escalation. Zero dependencies, nothing to compile. |
| RavenGPT | An AI chat client that runs entirely in the browser; your OpenRouter key, nothing logged. Live at raven-gpt0.vercel.app |
| BurmeseGLUE | A shared benchmark for Burmese NLP: four tasks, fixed splits, reproducible baselines. |
| PortSwigger Cheat Sheets | Payloads and working notes across thirty topics of the PortSwigger Web Security Academy. |
| CogniVault RAG Engine | A Next.js retrieval pipeline where the retrieved chunks travel with every answer. |
| KeyForge | Bulk API key validation behind a terminal-styled interface. |
Elsewhere here: GetVideo, a Firefox repack of Video DownloadHelper, and pafish, a mirror kept open for malware analysts.
Five certifications held (CEH, CySA+, BTL1, eCTHP, CRTA), three published CVEs, and a seventh-place finish at the Magnet Virtual Summit 2026 in digital forensics. James Cook University (Singapore Campus). Currently competing in national-level CTFs, including NCSA Thailand's AI CTF.
TypeScript (React/Next.js) for anything on the web, Python for research and scripting, Django when a backend needs to exist yesterday. Security work runs on Kali Linux; reverse engineering mostly happens in Ghidra.
Working through the PortSwigger Web Security Academy systematically and expanding the payload sheets. Deepening Burmese NLP work around BurmeseGLUE. Finishing the projects that have been sitting at around ninety percent done.