I reverse-engineer live threats, then build the OSINT / GEOINT and authorized-offensive tooling that answers them.
Founder & principal engineer · NO-DO · Desmonto amenazas reales y construyo lo que las responde.
eva420— the engineering home of NO-DO, a founder-led security & OSINT studio. One person covers the chain: recon, authorized operation, infrastructure, signed evidence, deploy on Cloudflare. Offensive and defensive work, always with written authorization.
salat-stealer-analysis is an independent, end-to-end reverse-engineering of an active Go-based Malware-as-a-Service infostealer (Salat Stealer / WebRat, sold by Russian-speaking actors). Starting from a trojanized "gamer tools" pack, I reversed the garble-obfuscated Go loader in Ghidra (headless), extracted and decrypted its self-embedded payload — ChaCha20-Poly1305 + zstd, CBOR-chunked — to recover a 12.5 MB PE32 stealer, detonated it in an isolated lab, and pulled the live C2 straight off the TLS ClientHello SNI after the sample hid resolution behind DNS-over-HTTPS. A purely passive urlscan.io pivot by IP then expanded the 5 documented C2 domains to 12 domains plus a live admin panel whose origin was exposed by an operator OPSEC failure. Attribution is kept group-level and sanitized — no victim data, no named individuals — and the write-up ships a full IOC table with a machine-readable iocs.csv.
Read the full write-up →
eva420/salat-stealer-analysis
What I work on and where you can check it. The code of the products is private; each row points to a public repo or a live site.
| Discipline | Where it's proven |
|---|---|
| Malware RE & threat intelligence | salat-stealer-analysis — Ghidra RE, payload decryption, live C2 discovery, published IOCs (public) |
| OSINT / GEOINT | TRACEZ — entity graphs, aircraft & ship watch, satellite imagery, GNSS integrity, timestamped evidence (in production) |
| Authorized offensive security | GRIMZ — scope-bound pentest tooling, live billing (in production) |
| Edge infrastructure & applied crypto | Cloudflare Workers + D1; Ed25519 node-locked licensing & signed evidence — applied crypto shown publicly in the salat payload decryption (ChaCha20-Poly1305) |
Research turned into tooling. Each product has its own site, so you can check it without access to a private repo.
| Product | What it is | Status |
|---|---|---|
| TRACEZ | OSINT/GEOINT workspace for investigations that cross records, territory and platforms: an entity graph with timeline and human-confirmed AI expansion over 36 open sources; a live map with 18 more (civil and military flights, AIS ships from a dedicated EU collector, wildfires, GNSS integrity, Sentinel-2/Landsat imagery); aircraft and ship watches with zone, loitering and AIS-gap alerts; plain-language questions to the map answered with cited, dated sources; web capture and Ed25519-signed reports with an RFC 3161 timestamp anyone can verify without an account; and an MCP server for AI agents. Cloudflare Workers + D1 + R2. Mobile Lighthouse 99 · CLS 0. | |
| GRIMZ | Authorized-pentest product: deny-by-default scope enforcement, node-locked licensing, enterprise engagements and controlled-intrusion training labs. | |
| ATTEST | Signs a finished case as a dossier a third party can verify without an account: Ed25519 signature plus an RFC 3161 timestamp, with a public verifier. |
How the work is kept inside its limits:
- Authorization-first. Offensive capability is gated on written scope and rules of engagement — enforced in code. GRIMZ ships deny-by-default scope: out-of-scope targets are refused by construction, not by convention.
- Scope-bound by construction. Licensing is node-locked (Ed25519) — tooling runs where it is authorized to run, and nowhere else.
- Evidence you can verify. Deliverables are Ed25519-signed, so a third party can check integrity and origin without having to trust me.
- Defender-focused. Public research ships at TLP:CLEAR with IOCs for blue teams; attribution stays group-level; labs are isolated and disposable.
- No rented offensive access. NO-DO deliberately does not sell hosted offensive capability — a liability line held on purpose, not a missing feature.
Coordinated disclosure & full policy: SECURITY.md · contacto@no-do.dev
TypeScript (strict) end to end · Python for tooling and analysis · Go for reverse-engineering and binary internals · Cloudflare Workers + D1 at the edge · Ed25519 for node-locked licensing and signed evidence · Ghidra and tshark for reverse-engineering · a local LLM for analysis, so investigation data never leaves the box.
Authorized security and OSINT engagements, research collaboration or hiring:
- Email — contacto@no-do.dev
- Web — no-do.dev
Open to collaboration and select security roles. · Abierto a colaboración y a proyectos seleccionados.
+ eva420 · NO-DO · recon → operation → signed evidence · Remote · EU


