Hacker News: waqarjaved

Framework choice explains ~0.06% of agentic AI security outcome (7,020 trials)

waqarjaved — Fri, 31 Jul 2026 23:17:35 +0000

Article URL: https://figshare.com/articles/preprint/Cross-Framework_Portability_of_Agentic_AI_Security_A_Controlled_Payload-Verified_Evaluation/33110642

Comments URL: https://news.ycombinator.com/item?id=49129575

Points: 3

# Comments: 0

Prompt Injection Is Not a Chatbot Problem: How the Attack Surface Changes

waqarjaved — Tue, 30 Jun 2026 17:54:41 +0000

Article URL: https://agentsafelabs.com/blog/prompt-injection-is-not-a-chatbot-problem-how-the-attack-surface-changes-when-your-llm-has-tools/

Comments URL: https://news.ycombinator.com/item?id=48736501

Points: 3

# Comments: 0

The OWASP Agentic Security Initiative Top: A Practical Developer Guide

waqarjaved — Tue, 30 Jun 2026 00:57:24 +0000

Article URL: https://agentsafelabs.com/blog/the-owasp-agentic-security-initiative-top-10-a-practical-developer-guide-for-langchain-and-crewai/

Comments URL: https://news.ycombinator.com/item?id=48727327

Points: 9

# Comments: 0

The OWASP Agentic Security Initiative Top: A Practical Developer Guide

waqarjaved — Thu, 25 Jun 2026 18:34:30 +0000

Article URL: https://agentsafelabs.com/blog/the-owasp-agentic-security-initiative-top-10-a-practical-developer-guide-for-langchain-and-crewai/

Comments URL: https://news.ycombinator.com/item?id=48677476

Points: 3

# Comments: 0

AgentSafeLabs – Launched Open-source Security framework for AI agents

waqarjaved — Wed, 27 May 2026 17:46:20 +0000

Article URL: https://github.com/AgentSafeLabs/safelabs-eval

Comments URL: https://news.ycombinator.com/item?id=48297782

Points: 1

# Comments: 0

New comment by waqarjaved in "Using AI to write better code more slowly"

waqarjaved — Tue, 26 May 2026 20:01:15 +0000

There's an easy workaround that helps instead of listing options, just describe the problem constraints and ask it to propose approaches independently.

New comment by waqarjaved in "Using AI to write better code more slowly"

waqarjaved — Tue, 26 May 2026 19:58:04 +0000

Yes agreed with you on this.

New comment by waqarjaved in "Using AI to write better code more slowly"

waqarjaved — Tue, 26 May 2026 19:54:41 +0000

Using AI to find bugs rather than write features is an underexplored mode. The incentives in most teams push toward generation speed, but "AI as the skeptic that reads your PR" has a much better quality to risk ratio for established codebases.

New comment by waqarjaved in "I spent 50 hours drawing a line graph"

waqarjaved — Mon, 25 May 2026 23:10:40 +0000

Hi, the irony of spending 50 hours on something specifically designed to communicate quickly is not lost on me. Would love to see a time-lapse.