jp101216
Flint, MI · Security Engineer | Offensive Security Research & Custom C2 Tooling · Independent
Houston County High School
7 followers
About
Cybersecurity-focused IT professional with 8+ years of enterprise support experience and hands-on exposure to SOC operations, blue team workflows, ethical hacking, and security testing. Experienced in incident response, log and telemetry analysis, vulnerability assessment, and documentation across Windows/Linux environments.
Specialties
- Penetration Testing
- Red Team Operations
- Security Engineering
- SOC Analyst
- Threat Detection
Skills
- Security Alert Monitoring
- Log Analysis
- SIEM Telemetry Analysis
- Threat Investigation
- Incident Response
- Indicators of Compromise (IOCs)
- Endpoint Security
- Forensic Triage
- Memory Integrity Verification
- Vulnerability Assessments
39 skills in total
Experience
-
Independent Security Researcher & Home Lab Engineer
Independent
Jan 2021 — Present
Home Lab/Remote
-
E-Commerce Manager / Lead Operations Clerk
Kroger
Jun 2009 — Sep 2018
Tennessee
-
Freight Flow Associate
Sam's Club
Jul 2026 — Present
Grand Blanc, MI
7 experiences in total
Education
-
Houston County High School
Diploma
Unknown — Unknown
Repositories
-
jp101216/Red-Team-Notes
Red Team and APT Emulation Notes and Reports
-
jp101216/LiveResponse-NG
Live-response triage agent for Windows and Linux DFIR: ~120 ATT&CK-mapped detections, memory image integrity (module stomping, hollowing), AI-aware malware detection, and an offline report viewer. Go + Rust.
Go
-
jp101216/Threat-Intelligence-Reports
Threat intelligence reports and APT profiles — adversary TTPs mapped to MITRE ATT&CK, with emulation and detection notes.
-
jp101216/Powershell-Ops-IR-Toolkit
A small suite of powershell tools. Endpoint triage, incident response, and day-to-day system maintenance. Built with a consistent house style, structured param blocks, color coded console output, a quiet mode that emits a ServiceNow pasteable output and Json support for every run.
PowerShell
Recent posts
Two security bugs I shipped in ITriage v0.10.0, and what fixed them in v0.10.1
ITriage is my open-source IT triage and repair tool (Rust, Windows and Linux). Technicians often run it elevated on a PC where a standard user is signed in, so anything that user controls is attack surface. A review before announcing 0.10.0 found two places where a new feature forgot that.
1. Privileged delete through a junction (CWE-59)
The new Clear Teams cache fix empties a folder in the console user's profile. Every folder on the way to it belongs to that user. Swap one for a junction, and an elevated ITriage would delete whatever it pointed at. The fix now walks each folder from the profile down and refuses if any is a link or reparse point.
2. Autostart from a path users can write to
Watch mode registers ITriage to start at every sign-in, as whoever signs in. Installed from a copy in Downloads or a folder at the root of C:, that file could be swapped, and the replacement would run as every user, administrators included. --install-watch now checks the file and every parent folder, and refuses unless only administrators can change them.
Both needed a local user plus a technician's elevated run, and I know of no one affected. Both have regression tests now.
The lesson: the right check already existed in the codebase both times. The temp cleanup refused links, and elevated runs already validated program paths. The bug was new code not calling them.
What 0.10 adds for defenders and the desk:
• Policy blocks: which ASR, AppLocker, WDAC, AppArmor or SELinux rule stopped a program
• Live CRL checks on machine certificates, before an expired CRL breaks 802.1X and VPN for everyone
• Kerberos tickets still encrypted with RC4
• Wired 802.1X failures and NAC quarantine
• Watch mode, and a banner when a run matches a known incident
Still untested: a real junction on desktop Windows, and a real clinical workstation.
Release notes: https://github.com/jp101216/ITriage/releases/tag/v0.10.1
(Screenshot is the app's sample data.)
LiveResponse-NG v0.19.0 is released: an open-source (MIT) live response and DFIR toolkit for Windows and Linux. It's a Go agent plus a Rust sidecar for YARA-X memory scanning.
This release starts with a full source review (agent, sidecar, viewer, build) and fixes for what it found. New detections:
• MCP servers whose env loads code: LD_PRELOAD, LD_AUDIT, NODE_OPTIONS --require/--import, JAVA_TOOL_OPTIONS -javaagent, BASH_ENV, .NET startup hooks and profilers, and temp dirs on PYTHONPATH/LD_LIBRARY_PATH. Configs from Continue, Goose, Amazon Q, VS Code Insiders, VSCodium and Cline/Roo in Cursor are now read too
• AES Kerberoasting: a burst of AES-encrypted tickets to many user-account services in 15 minutes
• DCSync by a machine account, against domain controllers inferred from which hosts wrote KDC and DS events
• Sigma process_creation rules on Security 4688: Image, ParentImage, User, IntegrityLevel and decimal PIDs are aliased from the 4688 fields
• Linux: all systemd Exec* keys, anacron, queued at jobs, bash -c "$(curl …)" and bash <(curl …), and privileged groups held as a primary GID (docker, disk, libvirt, incus-admin)
• zstd journald fields, read with a pure-Go RFC 8878 decoder (fuzzed, verified against libzstd and journalctl)
• Developer toolchain dirs are no longer flagged as hidden payloads
Binaries for Linux amd64/arm64 and Windows x64/x86/arm64, with SHA-256 checksums:
https://github.com/jp101216/LiveResponse-NG/releases/tag/v0.19.0
ITriage 0.9.0 is out.
Quick recap: ITriage is an open source IT triage, repair and update tool for Windows and Linux. Rust engine, Tauri desktop app, command line. It grew out of the scripts I used while on a healthcare IT service desk. It's MIT-licensed now.
WHAT I NEVER ANNOUNCED
• Security posture (pictured, with sample data): BitLocker on the system drive, antivirus and the age of its definitions, EDR sensors, the firewall (critical when it's off for the profile in use), Secure Boot, TPM, UAC, and who is an administrator of the PC. On Linux: LUKS, firewalld/ufw/nftables, and sudo and wheel members.
• Sign-in & account lockouts: is the account locked, disabled or about to expire, and what on this PC keeps locking it after a password change: saved credentials, a disconnected session, services and scheduled tasks running as that account, and failed sign-ins in the Security log. Saved credentials are listed by name only. The secrets are never read.
• BitLocker suspension: before a manufacturer's BIOS or firmware tool runs, BitLocker is suspended for one restart, so the PC doesn't come back asking for its recovery key. If protection is on and can't be suspended, nothing is installed.
• Hardening for when it runs elevated: Windows' own tools are started by full path, never by bare name. The manufacturer's update tool only runs from a folder standard users can't write to. A profile that a standard user dropped into ProgramData is ignored. On Linux, programs run as root come from the system's folders, not PATH.
• Likely cause: every report opens with what the evidence points to and why. Fixed rules, not a model, so the same findings always give the same answer.
BUGS WORTH ADMITTING
Two of these releases started as a review of the one before. Three things those reviews found:
• A one-click fix could start a real firmware run. After a fix, the tool re-checks the module the finding came from, and on the firmware module that re-check ran the install, even from a preview. A fix now never repeats a change.
• An opened report could write outside the history folder. The saved file's name used the report's own date untouched, so a crafted report could name another folder. Names are now built only from letters, digits and hyphens.
• The case bundle's privacy pass left Wi-Fi network names in: it looked for a field called "SSID", and the Wi-Fi check calls it "Network". It also shredded what wasn't private, such as system paths, version numbers and 8.8.8.8. Both directions are fixed in 0.9.0.
All three are in the release notes.
EVIDENCE FOR ESCALATION
• Change timeline: updates, installs, driver and service changes, restarts and Group Policy processing in plain sentences, placed against when the problem began. The report says a nearby change is a possible link, not proof.
• Recordings: CPU, memory, connection probes and Wi-Fi signal while the user reproduces the issue, with a marker for the moment it happens. No packet contents, screenshots or tokens.
• Entra & Microsoft 365 sign-in state (no tokens are read) and Wi-Fi/802.1X authentication history.
• Case bundles: a ZIP for the next tier with consistent pseudonyms, a preview of every file before export, opt-in logs and SHA-256 digests.
• New in 0.9.0: What's using the space, Startup & sign-in speed, and a guided "the whole PC is slow" workflow.
NOT TESTED YET
252 automated tests on Linux, 249 on Windows and 19 in a browser, and the checks run for real on CI runners. But nothing has run on a domain, as a standard user on Windows, or on a physical Windows workstation, and the fixes that change account state have never touched real state. The release notes list all of it. The Windows installers are still unsigned.
Try the interface in your browser (sample data, nothing to install): https://jp101216.github.io/ITriage/
Code, downloads and release notes: https://github.com/jp101216/ITriage
In the Lab: Building "Project Lattice" — A Next-Gen, AI-Integrated C2 Architecture in Go & Rust
Over the past few months, I’ve been heads-down designing and engineering a new open-source adversary emulation framework from the ground up: Project Lattice.
I wanted to share the architecture and feature set I’m currently building out to test how modern detection stacks hold up against bleeding-edge userland, kernel, and AI-assisted tradecraft.
Project Lattice pairs a containerized, microservice-based Go server architecture with "Wraith"—a lightweight, self-healing Rust implant engineered for deep in-memory stealth.
Architecture & Feature Set Currently in Development:
Containerized Microservice Server & Resilience (Go + Docker)
• Multi-Player Microservice Architecture: Fully Dockerized, Mythic-style decoupled services where listeners, payload builders, and browser-based operator consoles run as independent containers over an encrypted gRPC bus.
• Self-Healing Orchestration: Automated container watchdogs designed to recover faulted listeners or builder services and re-attach session state without dropping live implants.
• Graceful Server Shutdown: Signal-aware teardown that flushes encrypted campaign state, rotates ephemeral certificates, and safely hibernates active agents before container exit.
• Smart Transport Layer: Native HTTP/2, HTTP/3 (QUIC), DoH, mTLS, and WebSockets with automated SMB/TCP P2P mesh pivoting.
Dynamic Payload Generation & AI Obfuscation
• On-the-Fly Polymorphic Builds: Containerized build pipelines that compile unique agents per target on demand—stripping debug symbols, randomizing control flow, and scrambling import tables to defeat static hash and YARA detections.
• AI-Assisted Obfuscation & OPSEC: LLM integration via MCP to dynamically rewrite loader stubs, synthesize custom encryption routines, and lint operator commands against known EDR behavioral rules before execution.
• Environmental Keying: Cryptographic payload locking tied to target-specific domain/hardware attributes to neutralize automated sandbox detonation.
"Wraith" Implant (Rust): Bleeding-Edge Evasion & Kernel Tradecraft
• Silent Moonwalk: Full dynamic call stack desynchronization using ROP gadgets and synthetic unwind metadata to evade ETW stack-walking during indirect syscalls and sleep cycles.
• Advanced BYOVD Engine: Ring-0 research module for scrubbing kernel notify callbacks (Process/Thread/ObRegisterCallbacks), blinding ETW-Ti, and toggling PPL flags—complete with PatchGuard-safe kernel pointer restoration.
• In-Memory Stealth: Combining Module Stomping, Ekko/Zilean sleep encryption, Threadless Injection, and hardware-breakpoint (Dr0–Dr3) patchless AMSI/ETW blinding.
• Self-Healing & Zero-Trace Shutdown: Vectored Exception Handler (VEH) crash guards to prevent faulty BOFs from crashing the host process while auto-rerouting broken P2P links. On exit, Wraith restores all hooks, securely zeroes its own heap and image memory pages, and terminates with zero forensic footprint.
Everything is being built strictly for authorized red/purple team operations, adversary simulation, and detection engineering research.
I'd love to hear from the community: what evasion primitives, transport quirks, or AI workflows are you most interested in testing in your labs right now?
#RedTeam #AdversaryEmulation #C2 #RustLang #Golang #OffensiveSecurity #EDR #KernelResearch #AISecurity #PurpleTeam #BuildInPublic
DFIR and SOC folks, I'd like your help deciding what Live Response NG builds next.
Live Response NG (lrng) is my open-source live-response triage agent for Windows and Linux: a Go agent with a memory-safe Rust sidecar for the riskiest job, reading attacker-controlled memory. Today it has:
• About 240 named detections, each mapped to MITRE ATT&CK
• Windows memory analysis that catches module stomping, AMSI/ETW bypasses and sleep-masked beacons, and decodes Cobalt Strike and Havoc configs
• Active Directory attacks (DCSync, Kerberoasting, golden tickets, AD CS abuse) and BYOVD drivers checked against LOLDrivers
• AI-era coverage: malware that calls LLM APIs or carries embedded prompts, plus LLMjacking in cloud audit logs
• A one-click local console, an offline HTML report viewer and multi-host case workspaces
Reply with the number you'd use most:
1. Fleet-wide remote collection
2. SIEM and SOAR integrations
3. Signed, encrypted evidence packages
4. A printable findings summary
Using Velociraptor, KAPE, your EDR's live response or your own scripts instead? I want to hear from you too. What's the one thing you keep wishing your tool did?
If your idea ships, you'll get credit in the release notes.
https://github.com/jp101216/LiveResponse-NG
#DFIR #IncidentResponse #BlueTeam