← back
Stop babysitting your agents... — Brandon Waselnuk, Unblocked
Takeaway
Give agents curated, cross-system context before execution to reduce repeated human correction and structurally wrong implementations.
Summary
- Coding agents begin without organizational context, leaving engineers to repeatedly explain patterns, services, and business decisions.
- Unblocked's context engine combines static documentation with runtime signals and searches across corporate systems before returning a compact research packet.
- Naive retrieval can stop after the first plausible result, while adding MCP connections provides access without ensuring discovery or understanding.
- Larger context windows alone do not organize entities and relationships; code can compile successfully while violating the surrounding system's requirements.
context-engineeringunblockedcoding-agents
Original description
Same prompt. Same agent. Same model. Without a context engine: 2.5 hours, 20.9 million tokens, multiple rounds of human correction, and code that compiled but would have broken the entire system if it shipped. With one: 25 minutes, 10.8 million tokens, and a senior engineer who gave one nitpick and approved the merge. Brandon Waselnuk from Unblocked makes the case that the problem is not access but understanding. More MCPs give agents pipes to information. A million token context window just sits there. Naive RAG stops at the first result it finds, a phenomenon called satisfaction of search borrowed from radiology. What actually changes is a context engine that reasons across your codebase, Slack history, PR patterns, and org structure to build a research packet before the agent starts writing, so it arrives knowing your factory patterns, your fallback infrastructure, and what the CTO said was wrong in that thread three months ago. Speaker info: https://getunblocked.com