A web page that serves invisible text to detect LLM-powered crawlers. When an AI bot reads the hidden content and uses it in responses, the honeypot reveals the bot's behavior. It helps website owners identify unauthorized AI scraping.
Visit project ↗A deployed honeypot endpoint that logs every AI crawler or agent that requests it. It gives site owners visibility into which LLM-related bots are hitting their URLs. Likely useful for monitoring AI scraping behavior.
A deployed honeypot endpoint that logs every AI crawler or agent that requests it. It gives site owners visibility into which LLM-related bots are hitting their URLs. Likely useful for monitoring AI scraping behavior.
A developer released Honeyprompt, an open-source honeypot that balances load across multiple LLM providers and supports multiple protocols. The tool is designed to detect and log unauthorized LLM access attempts.
A developer released Honeyprompt, an open-source honeypot that balances load across multiple LLM providers and supports multiple protocols. The tool is designed to detect and log unauthorized LLM access attempts.
PromptTrace provides free, interactive labs where users can practice prompt injection and other LLM security attacks. It helps developers and security researchers understand and defend against LLM vulnerabilities through hands-on exercises.
PromptTrace provides free, interactive labs where users can practice prompt injection and other LLM security attacks. It helps developers and security researchers understand and defend against LLM vulnerabilities through hands-on exercises.
DeepSeek Harness link shared on Hacker News
DeepSeek Harness link shared on Hacker News
Codeberg.org published a blog post on July 23, 2026, discussing how large language models (LLMs) are being trained on free/libre open source software (FLOSS) code without proper attribution or compliance with licenses. The post explores potential measures to safeguard the FLOSS commons from such exploitation, sparking community discussion on Hacker News.
Codeberg.org published a blog post on July 23, 2026, discussing how large language models (LLMs) are being trained on free/libre open source software (FLOSS) code without proper attribution or compliance with licenses. The post explores potential measures to safeguard the FLOSS commons from such exploitation, sparking community discussion on Hacker News.