Neelagiri Chettiyar

About

Based in London. I work on security research and open-source tools, mostly around AI agents and the software they connect to. I build mcpgawk and gawk.dev. Always happy to talk shop.

Badges

Tastemaker
Tastemaker
Gone streaking 10
Gone streaking 10
Gone streaking
Gone streaking
Gone streaking 5
Gone streaking 5

Maker History

  • mcpgawk
    mcpgawkCatch MCP servers that change after you approved them
    Oct 2026
  • 🎉
    Joined Product HuntJune 19th, 2026

Forums

•

10d ago

Has an ai agent ever introduced a deprecated library into your project without you noticing?

Asked an agent to quickly add a web-scraping utility, and it wrote clean working code using an old npm package that hasn't been maintained in four years. The code worked perfectly in local testing, but security scanners flagged it the second we opened a pull request.

how do you restrict your development agents from introducing outdated or unvetted external dependencies?

•

10d ago

what is your rule of thumb for letting ai agents run git commands directly?

giving an agent terminal access to stage and commit changes speeds up prototyping, but it comes with weird surprises. Yesterday an agent staged temporary log files and untracked config edits into a feature branch commit without mentioning it in the summary.

do you give your agent full shell execution rights, or do you strictly lock down git operations to manual user approval?

•

10d ago

At what point during a long coding session do you notice your AI agent starting to hallucinate?

when starting a fresh thread, answers are crisp and precise. but after an hour of tweaking UI components back and forth, the agent starts referencing variables we deleted ten prompts ago.

I spent thirty minutes debugging an error only to realize the agent was silently modifying an outdated file copy it remembered from earlier in the session context.

how do you manage context drift do you hard-reset your agent chat regularly or prune system prompts?

View more