菜单
SkillsMP 已收集 strozynskiw/bonsai 中的 15 个 Skill。打开任一 Skill 可查看来源和详情。
已展示 15 / 15 个已收集 Skill。
Build, launch, and drive the bonsai TUI in an isolated PTY to verify behavior at the real surface — status bar, keybinds, slash commands. Use when verifying a TUI-facing change, or when /verify needs a TUI evidence-capture handle for this repo.
原文语言：英语
Use when adding or updating a built-in AI provider integration in the Bonsai Rust TUI agent, including researching official provider docs, listing models, deciding protocol support, modeling reasoning effort or variants, wiring auth/model metadata/registry…
原文语言：英语
Audit and refresh bonsai's builtin model catalog (models/builtin/*.toml) against models.dev and live provider listings — fix drifted context windows/output limits/pricing, add newly served models, retire stale ones. Use when a provider ships new models,…
原文语言：英语
Add, edit, troubleshoot, or explain any bonsai customization — custom agents, subagents and personas, themes, skills, MCP servers, hooks, verification/sandbox/read-isolation config, memory, steering files, providers, models, and local endpoints. Load before…
原文语言：英语
Go correctness and idiomatic style — the build/vet/test verify loop, module operations, error-handling discipline, naming and structure conventions, and the concurrency, slice, and interface pitfalls that produce wrong Go. Load before writing, reviewing, or…
原文语言：英语
Java correctness and idiomatic style — the Maven/Gradle verify loop, null-safety and Optional discipline, modern language features (records, sealed types, pattern matching), exception design, and clean API conventions. Load before writing, reviewing, or…
原文语言：英语
JavaScript correctness and idiomatic style — package-manager detection, the lint/test verify loop, ESM/CJS pitfalls, async and equality footguns, and clean modern conventions. Load before writing, reviewing, or refactoring JavaScript.
原文语言：英语
PHP correctness and idiomatic style — the Composer/PHPStan/test verify loop, strict types and typed properties, PSR standards, exception design, and clean API conventions. Load before writing, reviewing, or refactoring PHP.
原文语言：英语
Add or fix bonsai model providers and models — local OpenAI-compatible endpoints (Ollama, LM Studio, vLLM), API keys and authorization, new model entries with context windows, pricing, and reasoning options, or overrides of built-in catalog entries. Load…
原文语言：英语
Python correctness and idiomatic style — virtualenv/uv/poetry detection, the run-and-test verify loop, typing and lint tooling, Pythonic data modeling and naming, and the pitfalls that most often break Python edits. Load before writing, reviewing, or…
原文语言：英语
Ruby correctness and idiomatic style — the Bundler/RuboCop/test verify loop, Enumerable and block idioms, keyword arguments, exception design, and expressive conventions. Load before writing, reviewing, or refactoring Ruby.
原文语言：英语
Rust correctness and idiomatic style — the cargo verify loop, compiler-error recovery, ownership discipline, error-handling design, API and naming conventions. Load before writing, reviewing, or refactoring Rust code.
原文语言：英语
Shell scripting correctness and structure — dialect detection, strict-mode headers, quoting discipline, clean script layout, and the silent failure modes of pipes, subshells, and word splitting. Load before writing or editing shell scripts.
原文语言：英语
TypeScript correctness and idiomatic style — package-manager detection, the tsc/eslint/test verify loop, strictness and ESM/CJS pitfalls, type-driven design, and clean-code conventions. Load before writing, reviewing, or refactoring TypeScript.
原文语言：英语
HTML and CSS correctness and craft — semantic markup, accessible-by-default patterns, form behavior, nesting validity, layout discipline, and maintainable CSS architecture. Load before writing or editing HTML, CSS, or markup inside components and templates.
原文语言：英语