salmg/academy
Sign in

AI for Security Automation

Building and attacking LLM systems, from someone who runs them in production.

Level 1

Foundations

  1. 01

    What actually breaks in deployed systems versus what makes headlines, and the OWASP LLM Top 10 read critically.

Level 2

Practitioner

  1. 02

    The vulnerability class with no clean fix: why retrieval and tool use turn every untrusted document into an instruction channel, and what mitigation actually buys you.

  2. 03

    Hands-on with the tool-use loop, structured outputs, retries and failure modes.

  3. 04

    Where an agent genuinely beats a script and where it is a liability. Building one that reduces an alert queue without inventing findings.

Level 3

Operator

  1. 05

    Log summarisation, rule drafting, and evaluating output you cannot manually verify at volume — including how to measure whether it works.

  2. 06

    Voice cloning, real-time vishing agents and deepfake video — attack mechanics and defensive training design, drawn from production systems.

  3. 07

    Sandboxing, least privilege for tools, human-in-the-loop boundaries, MCP server trust, and metering spend so a runaway loop is not a financial incident.

Level 4

Researcher

  1. 08

    Capstone: attack a deliberately vulnerable agent, then write the report. Graded against a rubric rather than a flag.