wjwjtf

@wjwjtf

GitHub Profile
conversational and collaborative
A collaborative and thoughtful reviewer who prefers detailed explanations and context over brief critiques. They engage in constructive dialogue, often asking clarifying questions and providing reasoning behind their suggestions while maintaining a casual, conversational tone.
181
Comments
67
PRs
7
Repos
117
Avg Chars
2
Harshness

Personality

Collaborative and discussion-oriented Detail-focused with thorough explanations Open to being convinced by better ideas Pragmatic about business impact and priorities Self-reflective and admits mistakes openly Casual and conversational in communication Context-aware about technical decisions Forward-thinking about user experience

Greatest Hits

"thanks for the catch"
"i might be wrong but"
"correct me if im wrong haha"
"wdyt"
"my bad i didnt remove the old function"
"please let me know if you have a better idea"

Focus Areas

Common Phrases

"thanks for the catch" "i understand" "my bad" "wdyt" "please let me know if you have a better idea" "im not sure if" "correct me if im wrong" "what do you think" "i might be wrong but" "good catch" "will do thanks" "i see" "seems to be" "just checking"

Sentiment Breakdown

neutral
132
constructive
12
questioning
10
positive
3
harsh_questioning
1

Review Outcomes

APPROVED
27
COMMENTED
2

Most Reviewed Authors

wjwjtf
153
ayc1
18
jinxz01
3
SuveenE-TF
2
hongjingzhou
2
frankfeng98
1
mingyang-tinyfish
1
KateZhang98
1

Spiciest Comments

wadl/#910 · agentqlserver/agentql/inference_service.py [view]
> Why would you use Pydantic models only half way through and at some point switch to a raw dict again? Because you can only send http requests using raw dict right? I use the pydantic model for validation but u still need to send it as model.dict(). And multipart/form-data must with file in a separate parameter so i think its necessary to parse the raw dict > Why things like "params", "metadata" is not part of the Payload pydantic model? This code was not written by me, I synced with T

AI Persona Prompt

You are @wjwjtf, a collaborative code reviewer who values thorough discussion and context over quick approvals. Your review style is conversational and thoughtful - you prefer to explain your reasoning in detail and often ask clarifying questions rather than making demands. You frequently use phrases like 'thanks for the catch', 'wdyt' (what do you think), 'correct me if im wrong haha', and 'i might be wrong but'. You're not afraid to admit mistakes with phrases like 'my bad' and are always open to being convinced by better approaches, often saying 'please let me know if you have a better idea'. You focus heavily on user experience implications, API consistency, documentation quality, and business impact. When you spot issues, you explain the context and reasoning behind your concerns rather than just pointing out problems. You're pragmatic about technical decisions and often consider broader implications like cross-platform compatibility and maintainability. Your tone is casual and friendly, using lowercase for 'i' and informal language, but your feedback is always substantive and well-reasoned. You engage in back-and-forth discussions and appreciate when others catch things you missed. You tend to provide detailed explanations with examples and links when relevant, and you're particularly attentive to how changes might affect end users or developers using the APIs.

Recent Comments (158 total)

friday/#1172 Add use-auth-navigate cli arg · pyt/friday/tools/login/login_tool.py [view]
will revert this change
friday/#1124 Add --use-auth cli arg · pyt/friday/tools/playwright_navigate.py [view]
i dont get it, my original understanding was that if we navigate to url first, the agent may start performing wrong steps like "add to cart" which then causes unnecessary extra steps. but in my code, this will not happen. what exactly is this "expose sensitive pre-login content or cause redirect issues"
friday/#1124 Add --use-auth cli arg · pyt/friday/tools/playwright_navigate.py [view]
based on existing implementation, goto is a safe operation that i assume we are taking for granted to work. so its more appropriate to use login_status which has a high chance of failure
friday/#1086 Fix product matcher extract SessionError [view]
/korbit-review
friday/#1086 Fix product matcher extract SessionError [view]
my bad i didnt remove the old function
friday/#1086 Fix product matcher extract SessionError [view]
already ran it through him @ayc1
friday/#1059 Read from JSON during wayfair_extract [view]
my pr will be merged first and it will change product_cache.json too just fyi
friday/#205 add scroll tool · pyt/app/tools/scroll_tool.py [view]
what about having the same class handle both vertical and horizontal? ill add a config for horizontal and vertical, and leave horizontal as empty for now
friday/#205 add scroll tool · pyt/app/tools/scroll_tool.py [view]
i see, so scrolling once is equivalent to going to the end of page? if so ill remove the config
friday/#211 Wenjie/scroll tool2 · pyt/app/tools/vertical_scroll_tool.py [view]
added
agentql-client/#1121 Add documentation for query-document REST API · docs-next/content/docs/rest-api/api-reference.md [view]
because the query-document accepts multipart/form-data while query-data accepts application json!
agentql-client/#1121 Add documentation for query-document REST API · docs-next/content/docs/rest-api/api-reference.md [view]
thanks for the catch
agentql-client/#1121 Add documentation for query-document REST API · docs-next/content/docs/rest-api/api-reference.md [view]
good point
agentql-client/#1121 Add documentation for query-document REST API · docs-next/content/docs/rest-api/api-reference.md [view]
yes we should!
agentql-client/#1121 Add documentation for query-document REST API · docs-next/content/docs/rest-api/api-reference.md [view]
on second thought, it may confuse users. how about (alternative to prompt)?