Claude-subagent
Hvorfor denne subagent
De fleste RAG-demoer ser fine ud, indtil modellen svarer ud fra sin træningshukommelse i stedet for dine dokumenter. Den her baseline holder den ærlig.
Dokumenterne chunkes og embeddes i et Pinecone namespace; hvert spørgsmål henter de nærmeste chunks, og modellen svarer kun ud fra dem. Hvert svar citerer de chunks, det brugte, og når der ikke kommer noget relevant tilbage, siger den det i stedet for at digte et svar. Et rent udgangspunkt, du kan bygge videre på.
Sådan kører den
Used at step 01 to kick off the pipeline.
Used at step 01 to kick off the pipeline.
Used at step 01 to kick off the pipeline.
Used at step 01 to kick off the pipeline.
Eksempel på output
json// Sample output // (generated when the pipeline finishes) Answer questions from Pinecone-retrieved chunks only, citing the chunk IDs used and refusing when context is missing.
Unlock the rest
The full agent definition, install snippet, and starter task are gated for community members.
Members get the full `.md` agent file, the npm / pnpm install one-liners, a starter prompt that we've tuned against real runs, and the open-source repo when this automation ships there. One email, magic link, done.