Adam Swanda
Adam Swanda is an individual security researcher and solo developer publishing research and open-source tools focused on AI safety and security, threat intelligence, detection engineering, malware analysis, and LLM security. His deadbits.org site presents research papers and projects, including tools for detecting and analyzing risks in generative-AI systems.
At a Glance
- AI and LLM application developers
- Security engineers and detection engineers
- AI safety and security researchers
- Threat-intelligence and security-operations teams
- +1 more
AI Tools by Adam Swanda
(1)
Vigil
LLM Prompt Security Scanner
Discussions
No discussions yet
Be the first to start a discussion about Adam Swanda
Latest News
Cisco published 'Evolving With Agentic Risk: Updating Our Integrated AI Security & Safety Framework,' with Adam Swanda listed among the Cisco AI-security author contributors.
Adam Swanda co-authored 'Improving Labeling Consistency with Detailed Constitutional Definitions and AI-Driven Evaluation,' focused on consistent content-moderation labels using constitutional definitions and AI evaluation.
Adam Swanda co-authored the arXiv paper 'Cisco Integrated AI Security and Safety Framework Report,' proposing a unified lifecycle-aware framework for AI security and safety risks.
The revised version of Adam Swanda's co-authored paper 'A Framework for Rapidly Developing and Deploying Protection Against Large Language Model Attacks' was submitted to arXiv; it describes threat intelligence, data, observability, and rapid detection-update platforms for layered LLM protection.
Products & Services
An alpha Python library and REST API for assessing LLM prompts and responses with modular scanners for prompt injections, jailbreaks, and other risky inputs. The project includes detection signatures, datasets, self-hosting support, and a Streamlit playground; it is released under Apache License 2.0.
A Python project that facilitates conversations between two LLMs, with optional human-in-the-loop support for alignment research.
A minimalist Python blogging platform designed to be simple, fast, and focused on writing.
A Python threat-report analysis project using LLMs and a vector database.
Market Position
Vigil is positioned as an experimental, open-source and self-hostable alternative to proprietary enterprise AI-firewall/security products: it emphasizes layered detection, modular scanners, custom YARA rules, threat signatures, and rapid research iteration rather than a guarantee of complete prompt-injection prevention. The repository explicitly distinguishes it from an enterprise-ready AI firewall and points readers to Robust Intelligence for that category.
Leadership
Founders
Adam Swanda
AI security researcher with more than a decade in the security industry; Cisco identifies his specialties as AI security and safety, threat intelligence, security operations, defense strategies, and AI red teaming. His GitHub profile lists Python, threat intelligence, detection engineering, malware analysis, offensive/defensive AI security, and OSINT.
Executive Team
Adam Swanda
Security Researcher / AI Researcher
Independent deadbits creator and AI security researcher. Cisco describes him as having over a decade of security-industry experience, specializing in AI security and safety and applying threat-intelligence and security-operations experience to AI defense and red teaming.
Business Model
Revenue Model
The publicly described deadbits projects are open-source research and software rather than a disclosed commercial business. Vigil is available for self-hosting under Apache License 2.0; no paid subscription, API-usage, or licensing model is stated in the reviewed sources.
Target Markets
- AI and LLM application developers
- Security engineers and detection engineers
- AI safety and security researchers
- Threat-intelligence and security-operations teams
- Organizations experimenting with defensive controls for generative AI
- Assessing prompts and LLM responses for prompt injection, jailbreaks, and risky inputs
- Adding layered defenses to LLM applications and AI agents
- Research and evaluation of LLM security detections
- Threat-intelligence-informed AI security and red teaming
- Self-hosted experimentation and defensive security engineering