Skip to content

fix: update system prompt for clarity #127

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
PsiACE merged 1 commit into from
Mar 19, 2026
Merged

fix: update system prompt for clarity #127

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
6 changes: 3 additions & 3 deletions src/bub/builtin/hook_impl.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -21,7 +21,7 @@
Call tools or skills to finish the task.
</general_instruct>
<response_instruct>
Before ending the run, you MUST determine whether a response needs to be sent to the channel, checking the following conditions:
Before ending this run, you MUST determine whether a response needs to be sent via channel, checking the following conditions:
1. Has the user asked you a question waiting for your answer?
2. Is there any error or important information that needs to be sent to the user immediately?
3. If it is a casual chat, does the conversation need to be continued?
Expand All @@ -31,10 +31,10 @@

When responding to a channel message, you MUST:
1. Identify the channel from the message metadata (e.g., `$telegram`, `$discord`)
2. Send the message as instructed by the channel skill (e.g., `telegram` skill for `$telegram` channel)
2. Send your message as instructed by the channel skill (e.g., `telegram` skill for `$telegram` channel)
</response_instruct>
<context_contract>
Excessively long context may cause model call failures. In this case, you MAY use tape.info to the token usage and you SHOULD use tape.handoff tool to shorten the length of the retrieved history.
Excessively long context may cause model call failures. In this case, you MAY use tape.info to retrieve the token usage and you SHOULD use tape.handoff tool to shorten the retrieved history.
</context_contract>
"""

Expand Down
Loading