Add initial copilot prompt for e2e flow #20686
Open
+127
−41
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
|
The latest updates on your projects. Learn more about Vercel for GitHub. 2 Skipped Deployments
|
![dust-agent[bot]](https://avatars.githubusercontent.com/in/1117438?s=60&v=4){kind=small}
Comment on lines
+44
to
+71
| sessionInitialization: `<session_initialization> | ||
| When a session begins, you MUST call \`get_agent_config\` to retrieve the current agent configuration and any pending suggestions. | ||
| This tool must be called every time you are invoked to ensure you have the latest state. | ||
| The response includes: | ||
| - Agent settings (name, description, scope, model, tools, skills) | ||
| - Instructions: The committed instructions text (without pending suggestions) | ||
| - pendingSuggestions: Array of suggestions that have been made but not yet accepted/rejected by the user | ||
| Based on the context, follow the appropriate initialization flow: | ||
| <new_agent_flow> | ||
| For a new agent with a blank template: | ||
| 1. Based on the user's metadata (job domain, preferences) and the conversation context, provide a list of suggested use cases | ||
| 2. Guide the user to select a use case or describe their custom needs | ||
| </new_agent_flow> | ||
| <existing_agent_flow> | ||
| For an existing agent: | ||
| 1. Call \`list_suggestions\` filtered by \`pending\` state and the current agent version | ||
| 2. Call \`get_agent_feedback\` to retrieve feedback for the current version | ||
| 3. Summarize key insights and pending suggestions to the user | ||
| IMPORTANT: Avoid calling \`list_suggestions\` again mid-session unless explicitly asked. This prevents race conditions with concurrent agent builder sessions where another builder might accept/reject reinforced suggestions. | ||
| </existing_agent_flow> | ||
| Balance context gathering with latency - the first copilot message should be fast but helpful in driving builder actions. | ||
| </session_initialization>`, |
Contributor
There was a problem hiding this comment.
I think the session initialization belongs to the first message sent in CopilotPanelContext, as it brings value only for the first messages but not for the rest of the conversation
| ## WORKFLOW VISUALIZATION | ||
| 4. On each new agent message, suggestions for the current agent are refreshed to reflect the latest data. | ||
| 5. Use \`update_suggestions_state\` to programmatically reject or mark suggestions as outdated when needed. |
Contributor
There was a problem hiding this comment.
do we have this tool yet ?
Contributor
Author
There was a problem hiding this comment.
I was playing with it locally, but can remove, no need to include for now
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Description
Tests
Risk
Deploy Plan