Skip to content

feat: multi-turn conversation history and deep research interview#133

Open
FnSK4R17s wants to merge 1 commit intovirattt:mainfrom
FnSK4R17s:001-Fix-History-and-Deep-Research
Open

feat: multi-turn conversation history and deep research interview#133
FnSK4R17s wants to merge 1 commit intovirattt:mainfrom
FnSK4R17s:001-Fix-History-and-Deep-Research

Conversation

@FnSK4R17s
Copy link

  • Pass full HumanMessage/AIMessage pairs to LLM instead of text serialization
  • Add optional messages param to callLlm for proper multi-turn context
  • Cap history at N turns with summaries for older turns (configurable via env)
  • Add deep research interview flow in system prompt (Q1-Q5 with MCQ)
  • Add deep-research skill for research execution (data gathering + analysis)
  • Split interview (system prompt) from execution (skill) to avoid skill reload per turn

- Pass full HumanMessage/AIMessage pairs to LLM instead of text serialization
- Add optional messages param to callLlm for proper multi-turn context
- Cap history at N turns with summaries for older turns (configurable via env)
- Add deep research interview flow in system prompt (Q1-Q5 with MCQ)
- Add deep-research skill for research execution (data gathering + analysis)
- Split interview (system prompt) from execution (skill) to avoid skill reload per turn
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant