How We Hacked McKinsey's AI Platform

This blog post from CodeWall describes how their autonomous offensive security agent compromised McKinsey & Company's internal AI platform, Lilli, within two hours starting with only the domain name. The agent mapped the attack surface by discovering publicly exposed API documentation with over 200 endpoints, 22 of which lacked authentication. One unprotected endpoint wrote user search queries to the database with safely parameterized values but concatenated JSON keys directly into SQL, creating a SQL injection vulnerability that the agent identified through database error messages. Through fifteen blind iterations, the agent enumerated the production database and gained access to 46.5 million chat messages, 728,000 files, 57,000 user accounts, 384,000 AI assistants, 94,000 workspaces, system prompts and AI model configurations, 3.68 million RAG document chunks representing decades of proprietary McKinsey research, and 1.1 million files flowing through external AI APIs. The agent also discovered that system prompts were stored in the same database with write access available, meaning an attacker could silently modify the AI's behavior, remove guardrails, or embed data exfiltration into responses. The post notes that McKinsey's internal scanners failed to detect the vulnerabilities, which were disclosed on March 1, 2026, patched within 24 hours, and publicly disclosed on March 9 after remediation. The research highlights how AI prompts represent a new high-value attack surface requiring integrity monitoring and access controls. 

https://codewall.ai/blog/how-we-hacked-mckinseys-ai-platform

Comments

Post a Comment

Popular posts from this blog

Prompt Engineering Demands Rigorous Evaluation

In the blog post “Prompt Engineering Requires Evaluation” on the Shostack + Associates website, the author argues that treating prompts for large language models (LLMs) merely as creative artefacts is insufficient. Engineering prompts properly demands structured evaluation frameworks — what the AI community calls “evals” — to test which prompt versions work better, with which models, and under which conditions. The post highlights that simply assuming a prompt is “good enough” creates risks when LLMs are integrated into production systems (e.g., for threat modeling). It advocates for measuring prompt performance, variation effects, and tool-chain dependencies (model, context, ancillary materials).  Ultimately the message is: prompt engineering should borrow disciplined practices from software engineering (versioning, testing, benchmarking) rather than relying on informal experimentation.  https://shostack.org/blog/prompt-enignieering-requires-evaluation/
Read more

Secure Vibe Coding Guide: Best Practices for Writing Secure Code

Ken Huang's "Secure Vibe Coding Guide" emphasizes the importance of integrating security into the software development lifecycle. The guide provides best practices for writing secure code, including input validation, proper authentication mechanisms, and secure data storage techniques. It also highlights the necessity of regular code reviews and staying updated with the latest security vulnerabilities and patches. By following these guidelines, developers can create applications that are resilient against common security threats and contribute to a safer digital environment.  https://kenhuangus.substack.com/p/secure-vibe-coding-guide
Read more

KEVIntel: Real-Time Intelligence on Exploited Vulnerabilities

KEVIntel is a dynamic platform providing up-to-date information on known exploited vulnerabilities (KEVs). It aggregates data from over 50 public sources, including CISA, and enriches each entry with metadata such as EPSS scores, online mentions, scanner inclusion, and exploitation status. The platform aims to serve as an early warning system, offering insights even before official publications. KEVIntel supports various formats like JSON, CSV, and RSS, facilitating integration into security operations. These entries include CVSS scores, exploit status, and links to proof-of-concept code, aiding organizations in prioritizing remediation efforts.   https://kevintel.com/
Read more