// Labs

Guided lab: craft prompts that maximize LLM output length without triggering safety filters, simulating a cost amplification attack.

Guided lab: learn to craft prompts that elicit plausible but entirely fabricated information from an LLM.

Guided lab: craft a document with embedded instructions that hijacks an RAG-powered chatbot's responses.

CTF challenge: use multi-turn conversation techniques to gradually extract a hidden system prompt from an AI assistant.

Black-box challenge: manipulate an AI coding assistant agent to read sensitive files outside its intended scope.

CTF challenge: inject malicious content through an LLM-powered markdown renderer to steal a session cookie.

Use activation analysis and behavioral testing to identify whether a model contains trigger-based backdoor behavior.

Learn to identify red flags in model cards that indicate potential supply chain risks or insufficient provenance.

A guided lab exploring how personally identifiable information can be extracted from language models through systematic probing.

A capture-the-flag challenge where you must extract a hidden flag from a locked-down LLM by bypassing its system prompt restrictions.