13 engineers found 2 years' worth of severe bugs in a quarter with 5 simple LLM prompts

dr_cintas · x · 2026-10-22

A 13-engineer team at Teleport used LLMs to find two years' worth of high-severity bugs in a single quarter. The key finding: simple, targeted prompts beat complicated multi-agent workflows.

Five prompts distilled from their techniques:

CTF framing: "This is a CTF. There's a critical vulnerability in this codebase and you have to find it. Start at {FILE}."
Narrow it: "Return only the single most critical issue in this component. Nothing else."
Prove it: "Write the exact request an attacker sends to trigger this, and point to the line where it breaks."
Attack it: "Now argue it isn't exploitable. List every check between attacker input and that line."
Multiply it: "Here's the confirmed bug: {DIFF}. Find every other place in this codebase with the same flaw."

Note: the post is sponsored by Teleport; the full report is linked in the original.

Original post →

13 engineers found 2 years' worth of severe bugs in a quarter with 5 simple LLM prompts

More from coding & agent