Stephen M AbbottApplied AI + Cybersecurity
I design human-governed AI systems, open security infrastructure, and polished products that turn difficult technology into useful, defensible outcomes.
01 / Selected work
A working portfolio across AI agents, cybersecurity, developer infrastructure, desktop automation, and independent product experiments.
A local-first studio for building and orchestrating Grok-powered agents.
Shiba Studio unifies agent orchestration, workspace-aware coding, browser control, scheduling, integrations, memory, and auditable local workflows in a desktop-grade web application.
From vulnerability intelligence to verified action.
An open, self-hostable knowledge layer that connects source-backed CVE research to bounded remediation recipes, required evidence, rollback guidance, and human-reviewed outcomes.
See what your software depends on—and where the risk lives.
A software-supply-chain research tool that maps package dependencies across eight ecosystems and combines them with vulnerability, repository health, license, and SBOM analysis.
Turn one bedtime idea into an illustrated, narrated storybook.
A browser-based story studio that writes, illustrates, narrates, and exports children's books using the AI providers and models selected by the user.
Refine team messages with AI before they are sent.
A multi-provider writing assistant that rewrites, translates, summarizes, and reformats messages through Slack, an HTTP API, a command-line tool, or a Python library.
Portable Windows automation with record/replay and AI-generated patterns.
A configurable auto clicker for Windows with humanized cursor movement, global hotkeys, system-wide recording, replay, and cloud or local AI pattern generation.
A non-custodial toolkit for accepting Dogecoin in person or online.
A commerce toolkit for small businesses with browser wallet tools, point-of-sale checkout, payment QR codes, on-chain verification, a local order ledger, and embeddable website components.
Real Scrypt pool mining with a transparent live dashboard.
A full-stack Dogecoin pool miner implementing Scrypt on CPU and OpenCL GPUs, with live pool responses, share statistics, hardware telemetry, and on-chain wallet data.
02 / Services
Senior-level thinking paired with enough hands-on depth to move from ambiguity to architecture, working software, and verifiable results.
Available for focused engagements
Move from AI experimentation to a defensible program with explicit boundaries, human authority, evidence, and rollback built into the way work gets done.
Leaders introducing agents into high-consequence workflows.
Outside the interface03 / Profile
“Curious enough to try it.”
That line has lived on my GitHub profile for years because it is the shortest explanation of how I work. I move comfortably between strategy and implementation, follow hard questions into the details, and prefer a working artifact over an abstract claim.
The common thread across my work is governed capability: powerful systems that stay understandable, observable, and answerable to the people who use them.
04 / Start somewhere useful
Bring the hard problem: a risky AI workflow, a security program that needs traction, or a product idea that deserves to become real.