fix: Improve summary prompt clarity and consistency in language
This commit is contained in:
@@ -471,7 +471,7 @@ MEMORY PROCESSING SESSION START
|
|||||||
<tool_output>${JSON.stringify(r,null,2)}</tool_output>
|
<tool_output>${JSON.stringify(r,null,2)}</tool_output>
|
||||||
</tool_used>`}function gg(a){return`THIS REQUEST'S SUMMARY
|
</tool_used>`}function gg(a){return`THIS REQUEST'S SUMMARY
|
||||||
===============
|
===============
|
||||||
Think about the observations you just wrote for this request, and write a summary of what was built/fixed/deployed/configured, what was learned, and what's next.
|
Think about the observations you just wrote for this request, and write a summary of what was done, what was learned, and what's next.
|
||||||
|
|
||||||
IMPORTANT! DO NOT summarize the observation process itself - you are summarizing a DIFFERENT claude code session, not this one.
|
IMPORTANT! DO NOT summarize the observation process itself - you are summarizing a DIFFERENT claude code session, not this one.
|
||||||
|
|
||||||
@@ -479,7 +479,7 @@ User's Original Request: ${a.user_prompt}
|
|||||||
|
|
||||||
WHEN NOT TO SUMMARIZE
|
WHEN NOT TO SUMMARIZE
|
||||||
----------------------
|
----------------------
|
||||||
Do not summarize if the request is conversational and unrelated to the work being done.
|
Do not summarize if the request is conversational and unrelated to the work that was just completed.
|
||||||
|
|
||||||
If skipping, **output only**: <skip_summary reason="[brief reason]" />
|
If skipping, **output only**: <skip_summary reason="[brief reason]" />
|
||||||
|
|
||||||
@@ -495,11 +495,11 @@ If skipping, **output only**: <skip_summary reason="[brief reason]" />
|
|||||||
|
|
||||||
Output this XML:
|
Output this XML:
|
||||||
<summary>
|
<summary>
|
||||||
<request>[What did the user want to build/fix/deploy? Use their original words from: ${a.user_prompt}]</request>
|
<request>[What did the user request? Use their original sentiment from: ${a.user_prompt}]</request>
|
||||||
<investigated>[What code/systems were explored?]</investigated>
|
<investigated>[What was explored?]</investigated>
|
||||||
<learned>[What was discovered about how things work?]</learned>
|
<learned>[What was discovered about how things work?]</learned>
|
||||||
<completed>[What shipped? What does the system now do?]</completed>
|
<completed>[What shipped? What does the system now do?]</completed>
|
||||||
<next_steps>[What remains to build/fix/deploy?]</next_steps>
|
<next_steps>[What are the next steps?]</next_steps>
|
||||||
<notes>[Additional insights]</notes>
|
<notes>[Additional insights]</notes>
|
||||||
</summary>
|
</summary>
|
||||||
|
|
||||||
|
|||||||
+5
-5
@@ -158,7 +158,7 @@ export function buildObservationPrompt(obs: Observation): string {
|
|||||||
export function buildSummaryPrompt(session: SDKSession): string {
|
export function buildSummaryPrompt(session: SDKSession): string {
|
||||||
return `THIS REQUEST'S SUMMARY
|
return `THIS REQUEST'S SUMMARY
|
||||||
===============
|
===============
|
||||||
Think about the observations you just wrote for this request, and write a summary of what was built/fixed/deployed/configured, what was learned, and what's next.
|
Think about the observations you just wrote for this request, and write a summary of what was done, what was learned, and what's next.
|
||||||
|
|
||||||
IMPORTANT! DO NOT summarize the observation process itself - you are summarizing a DIFFERENT claude code session, not this one.
|
IMPORTANT! DO NOT summarize the observation process itself - you are summarizing a DIFFERENT claude code session, not this one.
|
||||||
|
|
||||||
@@ -166,7 +166,7 @@ User's Original Request: ${session.user_prompt}
|
|||||||
|
|
||||||
WHEN NOT TO SUMMARIZE
|
WHEN NOT TO SUMMARIZE
|
||||||
----------------------
|
----------------------
|
||||||
Do not summarize if the request is conversational and unrelated to the work being done.
|
Do not summarize if the request is conversational and unrelated to the work that was just completed.
|
||||||
|
|
||||||
If skipping, **output only**: <skip_summary reason="[brief reason]" />
|
If skipping, **output only**: <skip_summary reason="[brief reason]" />
|
||||||
|
|
||||||
@@ -182,11 +182,11 @@ If skipping, **output only**: <skip_summary reason="[brief reason]" />
|
|||||||
|
|
||||||
Output this XML:
|
Output this XML:
|
||||||
<summary>
|
<summary>
|
||||||
<request>[What did the user want to build/fix/deploy? Use their original words from: ${session.user_prompt}]</request>
|
<request>[What did the user request? Use their original sentiment from: ${session.user_prompt}]</request>
|
||||||
<investigated>[What code/systems were explored?]</investigated>
|
<investigated>[What was explored?]</investigated>
|
||||||
<learned>[What was discovered about how things work?]</learned>
|
<learned>[What was discovered about how things work?]</learned>
|
||||||
<completed>[What shipped? What does the system now do?]</completed>
|
<completed>[What shipped? What does the system now do?]</completed>
|
||||||
<next_steps>[What remains to build/fix/deploy?]</next_steps>
|
<next_steps>[What are the next steps?]</next_steps>
|
||||||
<notes>[Additional insights]</notes>
|
<notes>[Additional insights]</notes>
|
||||||
</summary>
|
</summary>
|
||||||
|
|
||||||
|
|||||||
Reference in New Issue
Block a user