promptexploit

$ find /posts -type f -- 4 entries

• OWASP
  owasp-genai-llm-top-10-simple-explanation.md

  a plain-language map of the GenAI Security Project and LLM Top 10
  2026-05 [read →]
• PI
  indirect-prompt-injection-101.md

  how untrusted content smuggles instructions into a model
  2026-05 [read →]
• DEF
  isolating-tool-output-in-agents.md

  patterns that stop agents from trusting their inputs blindly
  2026-04 [read →]
• RT
  building-a-jailbreak-eval-harness.md

  measuring model robustness with a repeatable test suite
  2026-03 [read →]