Skill Detail

Run provider-agnostic terminal coding sessions with FuXi

Use FuXi when an operator wants a terminal coding agent to edit code, run commands, drive tools, and verify changes with cost-aware routing across LLM providers.

Developer ToolsMulti-Framework
Developer Tools Multi-Framework Security Reviewed
⭐ 1.8k GitHub stars
COPY SKILL INSTRUCTIONS (OPTIONAL)
npx skills add agentskillexchange/skills --skill run-provider-agnostic-terminal-coding-sessions-with-fuxi Copy
Uses the third-party skills CLI, not an ASE-owned installer. Check your agent’s compatibility. This copies instructions; complete the upstream tool setup below separately.
At a glance
Tools required
FuXi CLI; Git; ripgrep; a local project workspace; an OpenAI-compatible, Gemini, Bedrock, Vertex, or FuXi OAuth-backed model configuration; optional MCP servers and tools.
Install & setup
Inspect the upstream installer for the target platform, then install FuXi from https://releases.fuxicode.com/bootstrap.sh on macOS/Linux or the documented Windows installer. Verify the environment with fuxi –version, fuxi doctor, and fuxi verify before running fuxi inside a project.
Author
fuxicodex
Publisher
Organization
Last updated
Aug 23, 2026
Quick brief

Use FuXi when an operator wants to run a local terminal coding-agent session against a repository with provider-agnostic model routing, built-in file and shell tools, MCP client support, resumable sessions, checkpoints, and command guardrails. The agent workflow is to install the FuXi binary, run fuxi doctor and fuxi verify, open a project with fuxi, then let the terminal agent read code, edit files, run commands, use tools, and verify the result inside the workspace.

How it works

What this skill actually does

Invoke this instead of using a hosted coding product directly when the user needs a self-contained terminal workflow with bring-your-own-key model configuration, cost-aware provider routing, local transcripts, local checkpoints, MCP tools, and shell-visible verification. It is especially relevant for operators comparing coding-agent behavior across OpenAI-compatible providers or running autonomous code tasks under local terminal controls.

The scope boundary is terminal coding-agent operation in a local project. This is not a generic LLM provider card, SDK listing, or broad framework entry; it covers the repeatable operator workflow for starting, checking, routing, supervising, and resuming FuXi coding sessions in a repository.