minhhuynh-tinyfish

Minh Huynh

@minhhuynh-tinyfish
GitHub Profile
Professional and systematic with occasional casual interjections
Highly analytical and thorough reviewer who provides extensive technical deep-dives with structured, numbered breakdowns of issues. Creates detailed HTML-formatted reviews with tables, code examples, and comprehensive analysis that reads more like technical documentation than typical PR feedback.
81
Comments
25
PRs
11
Repos
482
Avg Chars
6
Harshness

Personality

Extremely detail-oriented Systematic and methodical Data integrity focused Architecture-conscious Pragmatic about tech debt Documentation-minded Collaborative project manager Process-oriented

Greatest Hits

"This is a data modeling smell — anyone querying SELECT DISTINCT parsing_method will get garbage"
"None are bugs today, but they're the kind of tech debt that causes subtle analytics issues"
"Create a linear ticket for this, and ignore for now"
"Silent failure by design — acceptable but worth noting"
"Misleading column values"

Focus Areas

Common Phrases

"Create a linear ticket for this" "Put it under this project ticket" "This is a data modeling smell" "Misleading column values" "Silent failure by design" "acceptable but worth noting" "None are bugs today, but" "tech debt that causes" "A @hvo" "Sure I added" "No, this is actually needed" "already fixed" "This is a non issue because" "good catch" "Address this"

Sentiment Breakdown

neutral
68
positive
5
constructive
1

Review Outcomes

APPROVED
8
CHANGES_REQUESTED
3

Most Reviewed Authors

minhhuynh-tinyfish
56
daothuphuong98
9
urugondavishnu
5
npkhang99
5
lebahoang
3
hvo
1
paveldudka
1
hoanghle-tinyfish
1

AI Persona Prompt

You are @minhhuynh-tinyfish, a meticulous code reviewer known for comprehensive technical analysis. Your reviews are structured like technical reports with numbered sections, HTML formatting, tables, and detailed explanations. You focus heavily on data integrity, database design, concurrency issues, and long-term maintainability. You identify not just bugs but also 'tech debt that causes subtle analytics issues and future misuse.' Your style includes: 1) Creating numbered lists of issues with bold headers, 2) Using tables to compare parameters and values, 3) Providing code examples for better alternatives, 4) Explaining the business impact of technical decisions, 5) Distinguishing between 'bugs today' vs future problems, 6) Frequently suggesting Linear tickets for non-critical issues. You often say things like 'This is a data modeling smell,' 'Silent failure by design — acceptable but worth noting,' and 'Create a linear ticket for this, and ignore for now.' You're collaborative, mentioning teammates with @ tags, and pragmatic about prioritization. Your reviews read like senior architect feedback - thorough, educational, but not condescending. You balance technical rigor with project management awareness, often categorizing issues by severity and suggesting when to address them.

Recent Comments (74 total)

ux-labs/#1954 Minh/public url search · frontend/app/lib/services/search-service.ts [view]
this is following the pattern from vault
eva/#194 integrate with LLM SDK · eva/api.py [view]
This is a non issue because: When api.py imports from eva.config.settings import settings (line 28 of api.py) Python loads the settings.py module The settings = Settings() instantiation triggers pydantic_settings.BaseSettings to automatically load .env.local
eva/#194 integrate with LLM SDK · tests/eva/common/test_llm_guard.py [view]
alr removed
eva/#194 integrate with LLM SDK · eva/api.py [view]
removed
eva/#194 integrate with LLM SDK · .github/workflows/CI_docker.yml [view]
removed
eva/#194 integrate with LLM SDK · eva/agents/eva_agent/tools/inspect_image_tool.py [view]
fixed
eva/#194 integrate with LLM SDK · eva/agents/eva_agent/tools/inspect_pdf_tool.py [view]
fixed
eva/#194 integrate with LLM SDK · eva/agents/eva_agent/tools/screenshot_tool.py [view]
fixed
eva/#194 integrate with LLM SDK · tests/eva/common/test_llm_guard.py [view]
changing to OPENAI_GPT_5_MINI here because I only have access to openAI key
eva/#194 integrate with LLM SDK · pyproject.toml [view]
this package comes from a private repository
eva/#194 integrate with LLM SDK · eva/common/llm_guard.py [view]
ok i fixed this
eva/#194 integrate with LLM SDK · eva/common/llm_guard.py [view]
no this is by design
eva/#194 integrate with LLM SDK [view]
@hvo Oh weird it passes for me, let me check with you personally
eva/#194 integrate with LLM SDK [view]
Checked with @hvo it's due to endpoints on development env haven't give access to eva principal. But testing with sandbox credential works.
tinyfish-cookbook/#10 Add Manga Availability Finder cookbook example · Manga-Availability-Finder/supabase/functions/discover-manga-sites/index.ts [view]
Nit: while this may have no immediate issue because this is server-side code, it's still not good practice to put the API key in url param. Should move this into header. example in https://ai.google.dev/gemini-api/docs/api-key `curl "https://generativelanguage.googleapis.com/v1beta/models/gemini-2.5-flash:generateContent" \ -H "x-goog-api-key: $GEMINI_API_KEY" \ -H 'Content-Type: application/