SECURITYBUGFOCUS/SIM4N6 Book a Call
SecurityBugFocus · Independent Security Research · Tangier, MA · Est. MMXXII

I find the bugs vendors would rather not exist.

I'm Sim4n6 — a solo ethical hacker working under the SecurityBugFocus name. I run vulnerability research and bug bounty engagements against web applications, CI/CD pipelines, and modern AI/MCP stacks. 37 CVEs disclosed across GitLab, NVIDIA NeMo, DataDog, MindsDB, Plane, MobSF, Caido, and a long tail of open-source software since 2022.

DISCLOSURE LEDGER
CVEs disclosed37
Critical04
High10
Medium19
Low04
Years active2022 → today
§ 01

What I do for hire.

FIVE OFFERINGS
01

Application Security Audit

Manual pentest of your web app, API, or GraphQL endpoint — focused on the bug classes I actually find: authorization, input handling, parser confusion, and the edge cases your scanner missed.

Web AppsREST / GraphQLAuth FlowsOAuth / SSO
02

CI/CD & Supply-Chain Review

GitHub Actions, GitLab CI, build-pipeline integrity, secret-handling, and the long tail of injection paths through your release process. Where I've found a Critical command-execution before.

GitHub ActionsGitLab CIRelease PipelinesSecrets
03

Static Analysis Engagement

Custom CodeQL queries written for your codebase, with taint tracking from real attacker-controlled sources to the sinks that matter. Findings come with working PoCs, not theory.

CodeQLRuby / RailsPythonJavaScript
04

AI / MCP Security Review

Prompt injection, tool-call abuse, agent authorization, and MCP server boundary review — the security model nobody got around to specifying yet, looked at by someone who's actually built one.

LLM Red-TeamMCP ServersAgent AuthTool-Call Abuse
05

Disclosure Coordination & Bug-Bounty Triage

You found something — or someone reported it to you — and now you need help. I write the advisory, coordinate with vendors and CNAs, request the CVE, draft the patch notes, and run the disclosure timeline so it lands cleanly. I've shipped 37 of these. Yours will be number 38.

CVE / CNA CoordinationAdvisory WritingVendor CommsDisclosure TimelineHackerOne / Huntr
§ 02

The receipts — CVEs & advisories.

37 DISCLOSURES · 2022 → 2026
37
Total · vulnerabilities reported
04
Critical · severity
10
High · severity
19
Medium · severity
04
Low · severity
2026 // 02 disclosures
Identifier
Finding
Target
Severity
37
CVE-2026-1388
Regular-expression denial-of-service in merge-request handling — ReDoS via crafted reference
GitLab CE/EE
High
36
CVE-2026-1101
Denial-of-service in GraphQL SBOM API — inefficient processing of attacker input
GitLab CE/EE
Medium
2025 // 08 disclosures
Identifier
Finding
Target
Severity
35
CVE-2025-13690
Denial-of-service in webhook custom headers
GitLab CE/EE
Medium
34
CVE-2025-12576
Denial-of-service in webhook endpoint
GitLab CE/EE
Medium
33
CVE-2025-13335
Endless server-side redirection in wiki — crafted wiki file triggers infinite loop
GitLab CE/EE
Medium
32
CVE-2025-0673
Infinite redirect loop → denial-of-service
GitLab CE/EE
High
31
GHSA-6p2v-wcv8-8j6w
Arbitrary file read via "Copy as cURL" command flow
Caido Plugin
Critical
30
CVE-2025-0549
Partial bypass of device-OAuth flow via cross-window forgery
GitLab
Medium
29
CVE-2025-31116
SSRF via DNS rebinding in assetlinks_check
MobSF
Medium
2024 // 15 disclosures · selected
Identifier
Finding
Target
Severity
28
CVE-2024-47830
Server-side request forgery via /_next/image endpoint
Plane
Critical
27
CVE-2024-24759
Bypass of SSRF protection via DNS rebinding
MindsDB
Critical
26
CVE-2024-21623
Arbitrary expression injection in CI workflow → command execution & secret leak
OTClient
Critical
25
CVE-2024-0081
User-controlled Unicode filename triggers server-side DoS — Nvidia acknowledged
NVIDIA NeMo
High
24
CVE-2024-35231
DoS via unconstrained profiler_runs parameter
Rack::Contrib
High
23
CVE-2024-8124
DoS via large glm_source parameter
GitLab
High
2023 // 10 disclosures · selected
Identifier
Finding
Target
Severity
22
CVE-2023-30620
Arbitrary file write extracting remote tarball via tarfile.extractall()
MindsDB
High
21
CVE-2023-31131
Arbitrary file write when extracting tarballs via shutil.unpack_archive()
Greenplum DB
Medium
20
CVE-2023-35932
Configuration injection via unsanitised user input
jcvi
High
19
CVE-2023-25803
Directory traversal vulnerability
Roxy-WI
High
2022 // 03 disclosures
Identifier
Finding
Target
Severity
18
CVE-2022-1993
Path traversal in /info/refs endpoint
Gogs
High
17
CVE-2022-3607
ZipSlip symlink variant → arbitrary file read
OctoPrint
Medium
16
CVE-2022-23530
Arbitrary file write when scanning crafted PyPI package
GuardDog (DataDog)
Low

The complete list — every CVE, every advisory, with year-by-year detail and severity breakdown — lives in my GitHub profile and is updated as disclosures land.

↗   FULL LIST · github.com/sim4n6
§ 03

Field notes.

WRITING ON THE WORK
§ 04

Behind the handle.

ABOUT SIM4N6
[ ID · 0x534D34 ] SECURITYBUGFOCUS
SIM4N6
  • Handlesim4n6
  • BasedTangier · MA
  • HackerOne@sim4n6
  • GitHub/sim4n6
  • LanguagesEN · FR · AR
  • AvailableQ3 2026 onwards

SecurityBugFocus is a one-person practice. There's no agency behind me, no junior consultants quietly running the engagement. When you book me, you get me — the same person whose name is on those 37 advisories.

I've spent the last several years finding and disclosing vulnerabilities in software people actually use — GitLab, NVIDIA NeMo, DataDog tooling, MindsDB, Plane, MobSF, OctoPrint, Gogs, NVIDIA Frigate, and a long tail of open-source projects in the supply chain. Every disclosure has been responsible: vendor first, CVE assigned, patch shipped before public detail.

The work I take on is the work I'm good at: deep technical review of web applications, CI/CD pipelines, and the new AI/MCP attack surface. I write CodeQL queries, I read source, I run the exploit. If a finding can't be exploited under realistic conditions, I don't pad the report with it.

Specialisations · stack · what I touch
  • Path Traversal
  • SSRF · DNS Rebinding
  • Unicode & ReDoS DoS
  • Arbitrary File Write
  • CI/CD Security
  • OAuth / SSO Flows
  • GraphQL Authorization
  • Ruby / Rails
  • Python · Flask · Django
  • CodeQL
  • MCP / Agent Security
  • Caido / Burp

Book a call.

Engagements start with a 30-minute call to scope the work, agree the rules-of-engagement, and decide whether I'm the right person for the job. No pitch decks. Email below.

DIRECT — EMAIL ONLY

Response time: typically within 48 hours, Tangier working hours (UTC+01). For new engagements, please include a short description of the target, the desired scope, and your timeline. Bug-bounty advisory questions and disclosure-coordination requests welcome.