Skip to content
View fxia0018's full-sized avatar
👁️‍🗨️
Focusing
👁️‍🗨️
Focusing

Block or report fxia0018

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
fxia0018/README.md
natsuki — CS @ Monash '28, agents, MCP, RAG. 69 PRs opened, 20 merged upstream.

Projects

furima-mcp · MCP server for Japan's second-hand marketplaces. Searches Mercari, Yahoo! Auctions, Yahoo! Flea Market and Rakuma in one call, prices items against what actually sold (split by condition), and watches searches for new listings. Python, on PyPI and the official MCP Registry.

Merged upstream

Bugs I found, reproduced and fixed in other people's agent / RAG infrastructure. Updated daily.

Project What I fixed
langgenius/dify
★ 162k
#44837 clean away unnecessary __new__ in tests
Tencent/WeKnora
★ 38k
#5450 treat an unset MCP selection mode as none, not all
#5119 strip host context from getUserQuery for save-to-knowledge-base
#5057 send Authorization header on legacy MinerU /file_parse and /docs probes
agentscope-ai/agentscope
★ 36k
#4058 reject None metadata_filter values on stores that cannot match null
opensandbox-group/OpenSandbox
★ 18k
#2885 stop snapshot template tests from writing time.Local
#2842 set gin test mode once for controller tests
Tencent/AI-Infra-Guard
★ 7.6k
#783 make GitClone retries work after a timeout
#782 keep files extracted from .tgz targets readable
#781 close each extracted zip entry before the next
#777 keep severity of CVSS-score and Chinese finding levels
#776 honor finish tool name casing
#772 lock the callback slice in TestRunner_Callback_Invoked
#765 stop writing a package global on every log call
pydantic/logfire
★ 4.6k
#2714 materialize generator messages in instrument_anthropic
agentscope-ai/ReMe
★ 3.8k
#753 bracket IPv6 hosts in startup banner URLs
agentscope-ai/OpenJudge
★ 1.1k
#251 preserve tool call accuracy raw score before bounding
#250 keep texts shorter than n in n-gram Jaccard
#245 build ActionLoop signatures from dict and nested arguments
#244 ignore empty entries in contains_all/contains_any substrings

all of them →

Email

Pinned Loading

  1. furima-mcp furima-mcp Public

    MCP server for Japanese flea markets: search Mercari, Yahoo! Auctions, Yahoo! Flea Market and Rakuma, and check real sold prices

    Python

  2. Tencent/WeKnora Tencent/WeKnora Public

    Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.

    Go 37.9k 5.3k

  3. Tencent/AI-Infra-Guard Tencent/AI-Infra-Guard Public

    A full-stack AI Red Teaming platform securing AI ecosystems via Agent Scan, Skills Scan, MCP scan, AI Infra scan and LLM jailbreak evaluation.

    Python 7.6k 780

  4. langgenius/dify langgenius/dify Public

    Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without …

    Python 162k 25.5k

  5. agentscope-ai/agentscope agentscope-ai/agentscope Public

    Build and run agents you can see, understand and trust.

    Python 36.4k 3.9k

  6. pydantic/logfire pydantic/logfire Public

    AI observability platform for production LLM and agent systems.

    Python 4.6k 376