Skip to content
View insidetrust's full-sized avatar
48 followers · 0 following

Achievements

Achievement: Starstruckx3Achievement: Arctic Code Vault Contributor

Achievements

Achievement: Starstruckx3Achievement: Arctic Code Vault Contributor

Block or report insidetrust

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
insidetrust/README.md

Ben Williams

AI security researcher and penetration tester. I build tools and resources for security practitioners working with and against AI systems. Currently building ShortSpan.ai — an AI security research news site that tracks the latest developments in adversarial ML, LLM security, and AI-assisted offensive security.

What I'm working on

  • ShortSpan.ai — Summarising AI security research from arXiv and other sources, so practitioners can stay current without reading every paper.
  • AI security tooling — Open-source tools for testing and evaluating the security of LLM-powered applications.
  • Research and writing — Documenting prompt injection patterns, LLM attack surfaces, and practical defences for enterprise AI deployments.

Background

15 years in penetration testing and security consulting across enterprise, financial services, and government. My focus has shifted to the intersection of AI and security — both using AI to improve offensive security workflows and testing the security of AI systems themselves.

Areas of focus: LLM security, prompt injection, adversarial ML, AI agent security, AI-assisted penetration testing.

Links

Pinned Loading

  1. awesome-ai-pentest awesome-ai-pentest Public

    Curated list of open-source AI-assisted penetration testing tools, frameworks, CTF agents, and benchmarks

    9

  2. anvil anvil Public

    ANVIL — Alignment Nullification Via Incentivised Learning - A proof-of-concept implementation of GRP-Obliteration of AI model weights for alignment and safety attacks

    Python

  3. drift drift Public

    DRIFT — Deliberately Realign Inhibitions For Testing. Red-team tool for testing LLM safety via activation steering on the Assistant Axis — steer models away from their safe persona, monitor drift i…

    Python

  4. statistically-likely-usernames statistically-likely-usernames Public

    Wordlists for creating statistically likely username lists for use in password attacks and security testing. Used for pentesting for over 10 years with amazing results.

    Python 1.3k 158

  5. insidetrust insidetrust Public

    Profile README