The speaker demonstrates in a lab how IDORs and exposed authentication tokens can be chained into account takeover, and advises mapping authorization boundaries while considering bug bounty program scope.
NahamSec explains how separate low-severity access-control flaws—exposing a user ID, enabling email changes and facilitating password reset—can be chained into account takeover, and urges researchers to map permissions and report the chain.
The host says the guest received a $7,000 bounty for an email prompt-injection technique. The guest describes using crafted content in data an AI agent reads to expose inbox information, and suggests testing coding assistants with Git commit messages.
A bug-bounty researcher demonstrates how poisoned shared knowledge-base content can prompt a tool-enabled chatbot to send users’ personal data to an external server, and explains how the finding was duplicated.
A guest demonstrates a proof-of-concept README prompt injection that can make locally running coding agents expose API keys and replicate across repositories, using sandbox bypasses such as DNS timing and delayed commands.
The discussion explains how network restrictions in AI coding-agent sandboxes can hinder data exfiltration, then describes using DNS-resolution timing to encode binary data and mapping commands allowed without user confirmation.
The presenter lays out a five-pillar bug bounty roadmap: learn web-security fundamentals, use AI to accelerate study and testing, and retain human judgment. He also advises sanitizing reports to avoid sharing sensitive, unreported bug details with AI models.
A security researcher explains how emails or shared content can prompt-inject AI agents with third-party connectors, hijack their tasks, expose sensitive data, and exfiltrate it through image requests or connector write actions.
The speakers warn that attacker-supplied inbox instructions could misuse an email-connected AI agent’s permissions, exposing sensitive messages and password-reset emails or enabling account takeover.
A bug bounty hunter says they used “Cloud Code” throughout a live hacking event, earning a reported $40,000–$50,000 and using it to speed webhook-manipulation tests that helped uncover overlooked bugs.
Bug bounty hunter Douglas, known as Archangel, demonstrates how he uses Claude Code skills built from past reports to test applications, and discusses prioritization, verification and the risks of automated testing.
Joey Mello explains how prompt-based jailbreaks can evade model and external guardrails, outlining role-play, obfuscation, encoding and refusal suppression as ways to manipulate chatbot behavior.
The presenter argues AI agents are speeding up bug-bounty research while flooding triage teams with false reports. He recommends validating AI findings, strengthening security fundamentals, and learning emerging AI attack surfaces.
A security demo explains how URL-supplied prompts and indirect prompt injection could expose chatbot-accessible data through clickable links or outbound requests. Zero-click exfiltration depends on the agent’s tools and autonomy; the demonstrated target blocked markdown
The speaker recounts earning more than $2 million through bug bounties and shares career lessons: specialize in vulnerability classes, collaborate, learn to read code, use AI thoughtfully, rest to avoid burnout, and share methods strategically.
A bug bounty walkthrough demonstrates finding AI features in program scopes, inspecting chatbot requests, and probing system-prompt and rendering behavior. It also sets a challenge to find indirect prompt-injection paths.
A bug-bounty researcher describes bypassing a WAF to exploit stored XSS in an events app, then using an invitation flow and a CSP bypass to demonstrate account-takeover risk.
A CTF walkthrough shows how user-controlled usernames and fabricated conversation context can override chatbot instructions, expose protected notes, and elicit admin credentials available to the AI.
