[GUEST ACCESS MODE: Data is scrambled or limited to provide examples. Make requests using your API key to unlock full data. Check https://lunarcrush.ai/auth for authentication information.]

[@abhaybhargav](/creator/twitter/abhaybhargav)
"Claude Skills has just been released. Its a powerful new way to "teach" your LLM where you can specify certain skills in the form of markdown docs PDF docs images slide decks and more. All without the need for MCP Tools or function calling. This is great for individual productivity. But I see it as a powerful new attack surface. Remember: * Claude (or any LLM) can be run with Agents in the background of your LLM-enabled app or Agentic App. * This might have skills included as part of it. Skills that you're completely unaware of. * These skills might spit out disinformation malicious code bad"  
[X Link](https://x.com/abhaybhargav/status/1980680327283503292) [@abhaybhargav](/creator/x/abhaybhargav) 2025-10-21T16:59Z 6535 followers, XXX engagements


"Built a High-Level breakdown of MCP Security Risks Super stoked to be training at our unique program at @codeblue_jp in Tokyo and later at @BlackHatEvents BHEU"  
[X Link](https://x.com/abhaybhargav/status/1981042207550275929) [@abhaybhargav](/creator/x/abhaybhargav) 2025-10-22T16:57Z 6534 followers, XXX engagements


"If you have a Pro Plus account on AppSecEngineer we now have upgraded our LLM infrastructure to support more than 300+ models So our labs which used to predominantly be Ollama or OpenAI has supported for all models that you can possibly dream of"  
[X Link](https://x.com/abhaybhargav/status/1981406618722951586) [@abhaybhargav](/creator/x/abhaybhargav) 2025-10-23T17:05Z 6532 followers, XXX engagements