request: ability to specify the first few messages in a chat
Nobody has claimed this yet.
- Dominant language
- Python
- Stars
- 12.5k
- Forks
- 998
- Avg merge
- 3d 13h
- Merged PRs (30d)
- 10
Description
Hi,
I've been tinkering lately with using a meta prompt to generate better prompts.
The idea is roughly that you give python for fibonacci sequence to an llm that takes as system prompt You are a master of the education system. Given a student request, you know how to reformulate it to add details and phrasing so that the highly regarded teacher receiving the request will answer is perfectly. and as prompt python for fibonacci sequence. The result is Could you recommend some resources or provide guidance on learning how to use Python to implement the Fibonacci sequence algorithm? That can then be sent to the actual chat.
(the example is terrible but I think that with a good system prompt it can actually be awesome)
Obviously being a huge fan of llm my first idea was to use templates or pipes and voila.
Unfortunately, I can specify the system and prompt of the "master of the education system" but when I want to give it to the actual chat I can't because I don't think you can currently specify the first message of a chat message. So I end up specifying it to the system prompt and just typing "go" manually as the first message but it's quite inelegant.
So I think it would be incredible to be able to specify in the cli some prompt messages instead of just the system prompt. What do you think :)?
Contributor guide
First steps
- Read the whole issue, then the project's contributing guide.
- Comment on the issue to say you are picking it up — it saves two people doing the same work.
- Fork the repository and make your change on a branch.
- Open a pull request that references the issue number.
Research direction
Start by locating the CLI code that handles the system prompt and the initial chat prompt, then inspect how messages are assembled before a request is sent. Clarify the intended command-line shape for supplying multiple initial messages and verify that the resulting conversation preserves their order and roles.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- python
- Domain
- ai, cli
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100