How Dropbox uses MCP and Dash to close the design-to-code security gap

Summary

Dropbox developed a system using the Model Context Protocol (MCP) and Dash's semantic search to bridge the gap between security threat models and code implementation. By retrieving original security documents during pull requests, an LLM agent automatically evaluates whether the proposed code adheres to previously agreed-upon security requirements. This approach surfaces design regressions and missing controls that traditional static analysis tools miss.

Takeaways

• Only 12% of pull requests explicitly link back to their original security design reviews, and over half are opened more than a month after the review occurred.
• Using semantic search via Dash's MCP server, Dropbox successfully mapped 80% of design reviews to implementing code, with 69% of connections only discoverable through semantic matching.
• LLMs integrated directly into code review workflows can compare implementation against design intent, catching security gaps even when code is functionally correct.

Related reading

• Dropbox · Mar 17, 2026

  How we optimized Dash's relevance judge with DSPy

  Dropbox Dash optimized its LLM-as-a-judge relevance scoring system using DSPy to migrate from expensive proprietary models to cheaper open-weight alternatives. By establishing automated feedback loops based on human agreement and strict JSON format validation, the team systematically generated robust prompts for new models. This reduced human-score disagreement by 45% and slashed model adaptation time from weeks to days while enabling 10x to 100x more data labeling.

  Ilya Yakovlev,Andrew Cheung,Binoy Dash,Simran Jumani,Dmitriy Meyerzon,Mark Breitenbach,Ishan Mishra,Kazuaki Okumura,Mike White,Kevin Altschuler,Facundo Agriel,Ishan Mishra,Eric Wang,Dmitriy Meyerzon

• Cloudflare · Aug 14, 2026

  How Cloudflare detects MCP traffic and helps secure it

  Traditional security boundaries assume human speed and judgment, but AI agents execute nondeterministic actions rapidly and can invoke unapproved Model Context Protocol servers through standard HTTPS traffic. To mitigate this shadow traffic, Cloudflare One introduces detection and inspection capabilities across clients, network boundaries, and MCP servers. Cloudflare Gateway leverages protocol-level headers and JSON-RPC inspection to identify direct MCP connections, verify managed paths, and enforce access through approved MCP Server Portals. Furthermore, Cloudflare Agents SDK v0.20.0 implements the stateless 2026-07-28 MCP specification while providing fallback handshakes for legacy servers. Organizations can combine Gateway traffic source rules with server middleware like WriteGuard to block unauthorized tool invocations before execution.

  AJ Gerstenhaber

• Dropbox · Feb 26, 2026

  Using LLMs to amplify human labeling and improve Dash search relevance

  Dropbox Dash uses large language models (LLMs) to amplify human labeling efforts for training its search relevance and ranking models. By validating and optimizing LLM evaluators against a small set of human-labeled internal data, Dropbox creates massive, high-quality training datasets offline for production rankers like XGBoost without incurring high latency or latency costs at query time.

  Ilya Yakovlev,Andrew Cheung,Binoy Dash,Simran Jumani,Dmitriy Meyerzon,Mark Breitenbach,Ishan Mishra,Kazuaki Okumura,Mike White,Kevin Altschuler,Facundo Agriel,Ishan Mishra,Eric Wang,Dmitriy Meyerzon,Dmitriy Meyerzon

• Dropbox · Mar 25, 2026

  Reducing our monorepo size to improve developer velocity

  Dropbox reduced its core server monorepo size from 87GB to 20GB after approaching GitHub's 100GB limit and suffering from hour-long clone times. The excessive bloat was traced to Git's delta compression heuristics interacting poorly with internationalization directory paths, generating inefficient diffs across different languages. By collaborating with GitHub to run an aggressive server-side repack with tuned window and depth settings, Dropbox cut clone times to under 15 minutes.

  Ilya Yakovlev,Andrew Cheung,Binoy Dash,Simran Jumani,Dmitriy Meyerzon,Mark Breitenbach,Ishan Mishra,Kazuaki Okumura,Mike White,Kevin Altschuler,Facundo Agriel,Ishan Mishra