Menu
Repository-level view of 12 collected skills across 8 GitHub repositories.
Top repositories by collected skill count, with their share of this creator catalog and occupation spread.
Diagnose if local LLM API servers (Ollama, LM Studio, llama.cpp) and python environments are running and configured properly.
Verify that generated hands-on code scripts run successfully and interface with local LLMs as expected.
Guide participants through a 3-phase math explanation agent build: terminal single-agent, terminal subagent orchestration, then backend plus UI prototype with text and image input.
Maintain this repository as a raw-sources plus persistent-wiki workspace. Use when ingesting docs into wiki, answering from wiki, or running wiki lint.
Search arXiv papers by keyword, author, category, and date, then return ranked results with links, abstracts, and citation metadata. Use when a user asks for arXiv paper discovery, latest papers, related work curation, author/category monitoring, or Korean prompts such as "arxiv 논문 검색", "최신 arxiv 논문", "arxiv 추천", and "관련 연구 찾아줘".
Maintain and extend the flutter_kiwi_nlp plugin. Use when working in this repository to implement features or fixes across native FFI and web WASM backends, adjust model loading behavior, update platform build scripts, and validate analyzer API parity before release.
Assess the threat of the current work and set the right MOPP protective posture before risky operations. Use before deploys, migrations, destructive commands, or touching auth/crypto/secrets/production — or whenever the user asks "what posture should I be in" or wants guardrails matched to the task.
Explains the MOPP (Mission-Oriented Protective Posture) model for AI work — the 0-4 levels, the cumulative gear/control stack, the dev and security faces, decision authority, and why de-escalation matters. Use when the user asks what MOPP is, how the levels map to guardrails, or which controls apply at a given level.