DEV Community

Royal Simpson Pinto profile picture

Royal Simpson Pinto

404 bio not found

Joined Joined on  github website twitter website
Why inbox AI only drafts, and what it takes to let it act

Why inbox AI only drafts, and what it takes to let it act

Comments
4 min read
MCP in practice: building, securing, and serving Model Context Protocol servers

MCP in practice: building, securing, and serving Model Context Protocol servers

Comments
5 min read
Building a streaming investigation assistant that tool-calls an MCP server and cites case IDs

Building a streaming investigation assistant that tool-calls an MCP server and cites case IDs

Comments
5 min read
Making a knowledge base queryable by any agent over MCP

Making a knowledge base queryable by any agent over MCP

Comments
4 min read
Scoping MCP tool access per client, and auditing every call

Scoping MCP tool access per client, and auditing every call

Comments
4 min read
Catching silent agent regressions in CI before they reach users

Catching silent agent regressions in CI before they reach users

Comments
4 min read
Put the Security Check Inside the Query, Not After It

Put the Security Check Inside the Query, Not After It

Comments
5 min read
Profiling an AI agent's context window: where the tokens actually go

Profiling an AI agent's context window: where the tokens actually go

1
Comments 5
4 min read
Learning prompt injection by attacking a deliberately vulnerable AI

Learning prompt injection by attacking a deliberately vulnerable AI

Comments 1
5 min read
Signed, verifiable receipts for RAG answers: what they actually prove

Signed, verifiable receipts for RAG answers: what they actually prove

Comments
5 min read
A perfect transcript can still be a wrong call: evaluating voice agents

A perfect transcript can still be a wrong call: evaluating voice agents

Comments 1
4 min read
Fail the build when your prompt gets dumber: evalgate for prompt regression CI

Fail the build when your prompt gets dumber: evalgate for prompt regression CI

Comments 1
4 min read
I built agentrace to catch the subagent runs I should not trust

I built agentrace to catch the subagent runs I should not trust

Comments 1
4 min read
I built a security linter for MCP servers, because nobody audits the tools we hand our agents

I built a security linter for MCP servers, because nobody audits the tools we hand our agents

1
Comments 1
4 min read
Enforcing RAG access control inside the retrieval query, not after it

Enforcing RAG access control inside the retrieval query, not after it

1
Comments 1
4 min read
I shipped five AI-infrastructure tools in a week, here is the throughline

I shipped five AI-infrastructure tools in a week, here is the throughline

Comments
3 min read
loading...