Voltar às skills
cli
por ByteAsk
Codex—Ainda sem avaliações
Sua avaliação
Resumo
AI coding agent for your terminal, with C/C++ build, debug, sanitizer and profiling tools. A fork of OpenAI Codex CLI.
Descrição
<h1 align="center">ByteAsk CLI</h1> <p align="center">An AI coding agent for your terminal, with tools for C and C++ work: build, debug, sanitize and profile.</p> <p align="center"> <a href="https://github.com/ByteAsk/cli/releases/latest"><img alt="Latest release" src="https://img.shields.io/github/v/release/ByteAsk/cli"></a> <a href="LICENSE"><img alt="License: Apache-2.0" src="https://img.shields.io/badge/license-Apache--2.0-blue"></a> </p> ByteAsk CLI is a fork of [OpenAI Codex CLI](https://github.com/openai/codex). It keeps Codex's terminal agent and adds native tools for C and C++ projects, a hosted model gateway with sign-in by email, support for your own provider keys and self-hosted models, and an optional connection to ByteAsk Context, a hosted code index. ByteAsk is not affiliated with, sponsored by, or endorsed by OpenAI. See [UPSTREAM.md](UPSTREAM.md). ## Install **macOS and Linux** ```sh curl -fsSL https://code.byteask.ai/install.sh | sh ``` **Windows (PowerShell)** ```powershell irm https://code.byteask.ai/install.ps1 | iex ``` **pip** ```sh pip install byteask ``` The installers download the engine for your platform from this repository's [releases](https://github.com/ByteAsk/cli/releases/latest) and check it against the published `SHA256SUMS`. Supported platforms: Linux x86_64 and arm64, macOS on Apple silicon, Windows x86_64. Release binaries up to v0.2.4 were built before the source was published; this repository holds that source, cleaned for publication. To build from source, see [docs/byteask/building.md](docs/byteask/building.md). ## Quick start ```sh cd your-project byteask ``` On first run ByteAsk asks how you want to run models, before it asks for anything else: - **ByteAsk account:** sign in with your email address; models run through the ByteAsk gateway. It is the same account as [app.byteask.ai](https://app.byteask.ai), where you manage your plan. - **Your own provider key:** OpenAI, Anthropic or Gemini; requests go from your machine to the provider. No account. - **Your ChatGPT plan:** sign in with a ChatGPT Plus or Pro subscription; OpenAI models only. No account. - **Your own hosted model:** any OpenAI-compatible endpoint, for example vLLM or Ollama. No account. Then describe the task. ByteAsk reads the code, runs commands in a sandbox, and shows each change for approval. ## Usage ```text byteask start an interactive session byteask "<prompt>" start with a prompt byteask exec "<prompt>" one answer, no UI, for scripts and CI byteask review review the working tree byteask resume | fork continue an earlier session byteask login | logout sign in or out byteask byok <set|status|remove> use your own provider key byteask models <add|test|list> use your own hosted model byteask context connect connect ByteAsk Context byteask mcp <add|list|remove> manage MCP servers byteask doctor check the install byteask --update update ByteAsk ``` Common options: `-m <model>`, `--effort <low|medium|high|xhigh>`, `-C <dir>`. Run `byteask --help` for the full list, and see [docs/byteask/usage.md](docs/byteask/usage.md) for examples. ## C and C++ tools When ByteAsk detects a C or C++ project it adds tools the agent can call on your code, for example: - build and test with the project's own build system, and select the tests a change affects; - run clang-tidy and cppcheck, and read the compilation database; - run AddressSanitizer, UndefinedBehaviorSanitizer and ThreadSanitizer builds; - debug with gdb or lldb, including core files, and join a gdb session you already have open; - record and replay with rr on Linux; - profile with perf and run benchmarks with noise checks; - show struct layout and padding, symbolize stack traces, and look up compiler output on Compiler Explorer. Each tool reports the command it ran and what it found. ## Configuration Settings live in `~/.byteask/config.toml`. The format follows Codex CLI's configuration, with ByteAsk's additions documented in [docs/byteask/configuration.md](docs/byteask/configuration.md). ## What ByteAsk sends - **Signed in with a ByteAsk account:** your requests, including the code context the agent sends, go to the ByteAsk gateway, which forwards them to the model provider. - **Your own key or your own hosted model:** requests go from your machine to that provider. ByteAsk receives a usage record after each turn (provider, model and token counts; never the key or the prompt). - **Install:** the installer reports the platform it installed for. - **Update check:** the launcher asks `code.byteask.ai` for the latest version. The launcher in this repository honours `BYTEASK_NO_TELEMETRY=1`, which turns off the install report and the usage record; it reaches installed copies with the next release. The [privacy policy](https://byteask.ai/privacy) covers what the hosted service does with requests. ## Relationship to Codex CLI The engine under `codex-rs/` is Codex CLI's Rust workspace. ByteAsk's own code lives mostly under `codex-rs/ext/`, plus changes to upstream files, all listed in [UPSTREAM.md](UPSTREAM.md). The launcher (`byteask`, `byteask.ps1`) and installers (`install.sh`, `install.ps1`) are ByteAsk's. We merge upstream stable releases regularly. The directories `codex-cli/`, `sdk/` and the upstream `docs/` are kept so merges stay small; nothing is published from them. ## Contributing Bug reports and pull requests are welcome. Read [CONTRIBUTING.md](CONTRIBUTING.md) first; everyone taking part follows the [Code of Conduct](CODE_OF_CONDUCT.md). Report security problems privately as described in [SECURITY.md](SECURITY.md). ## License Apache License 2.0. See [LICENSE](LICENSE) and [NOTICE](NOTICE). Third-party components are listed in [THIRD_PARTY_NOTICES.md](THIRD_PARTY_NOTICES.md).
Instalação
Instale a skill a partir de https://github.com/ByteAsk/cliEsta skill não está publicada no registro do ClawHub. Cole o prompt no seu agente de código — Claude Code, Cursor, Codex — e ele instalará a skill direto do repositório de origem.
Abrir o repositório