Agentic Search for Context Engineering
summary
Claim: good context engineering is a search problem, not just retrieval. Demos show a semantic-vector tool returning top-k hits, a general-purpose execute ESQL queries tool that initially fails on SQL wildcards (percent vs asterisk) until an ESQL agent skill injects correct syntax and error-handling, and a shell tool + Gina Grap flow that navigates files, runs grep-like queries, and finds the right session. Recommendations: log behavior, start general-purpose, then add low-floor specialized tools; the example stack works but is brittle without careful tool descriptions and retries.
AI-written summary · about
chapters (14)
- 0:00Introduction and Welcome
- 0:51Defining Context Engineering and the role of Search
- 2:21Historical context: From RAG to Agentic RAG
- 4:30Context sources (local files, memory, databases, web)
- 6:30Introduction to the Shell tool and its versatility
- 8:50Failure modes in agentic search
- 10:41The importance of tool descriptions and parameter design
- 13:53Code Demo: Simple semantic search and its limitations
- 23:26Code Demo: General purpose database query (ESQL)
- 28:36Code Demo: Adding Agent Skills for better interaction
- 34:42Code Demo: Using the Shell tool for file system retrieval
- 41:26Code Demo: Integrating custom CLIs (Gina Grap)
- 44:42Practical recommendations for building a search tool stack
- 49:16Q&A Session begins
related talks
31mBuild Intelligent AI Agents With Context Engineering
15mUnderstanding Context Engineering for Agentic Systems
31mBuilding the Context Engine AI Agents Need
12mUnconference: Context is all you need?
18mContext Engineering — Reliable AI Instead of Hallucinations
57mCrafting Intelligent Agents with Context Engineering
36mBeyond Grep: Search for Reliable Coding Agents
45mNever Send a Human to do an Agent's Search