flagos-ai/flagrelease/flagos-component-install

Name: flagos-component-install
Author: flagos-ai

Claude Code and agent skill by flagos-ai/flagrelease

Trust tier 3

community

By flagos-ai
AI/ML
flagos-ai/flagrelease
skills/flagos-component-install/SKILL.md
Updated 18 Oct 2026
Version 1.0.01 version recorded

Source check pending

Checking that the SKILL.md file is still in the repository.

FlagOS 生态组件统一安装/升级/卸载，支持 FlagGems、FlagTree

Install

npx vskill@latest install flagos-ai/flagrelease/flagos-component-install

SKILL.md

组件安装 Skill

通过 install_component.py 统一管理 FlagOS 生态组件的安装/升级/卸载。

工具脚本（已由 setup_workspace.sh 部署到容器）：

install_component.py — 统一入口
install_flagtree.sh — FlagTree 专用安装脚本（由 install_component.py 内部调用）

支持组件：

组件	安装方式	说明
`flaggems`	pip install（默认最新版）	FlagGems 算子库
`flagtree`	pip wheel / 源码编译（委托 install_flagtree.sh）	统一 Triton 编译器

上下文集成

从容器内 /flagos-workspace/shared/context.yaml 读取

yaml

container:
  name: <来自 container-preparation>
execution:
  cmd_prefix: <来自 pre-service-inspection>
inspection:
  flag_packages: <来自 pre-service-inspection>
gpu:
  vendor: <来自 container-preparation，FlagTree 安装时用于自动选择后端>

写入容器内 /flagos-workspace/shared/context.yaml

yaml

component_install:
  component: "<flaggems|flagtree>"
  action: "<install|uninstall|upgrade>"
  previous_version: "<old>"
  current_version: "<new>"
  install_method: "<pip|flagtree_pip|flagtree_source>"
  success: true|false
  timestamp: "<ISO>"
 
# FlagTree 安装后同步更新
flagtree:
  installed: true|false
  version: "<version>"
  triton_version: "<version>"
  install_method: "pip|source"
  backend: "<vendor>"
environment:
  has_flagtree: true|false

工作流程

步骤 1 — 查看当前版本

bash

docker exec $CONTAINER bash -c "PATH=/opt/conda/bin:\$PATH pip show flag-gems 2>/dev/null"

FlagTree 状态：

bash

docker exec $CONTAINER bash -c "PATH=/opt/conda/bin:\$PATH python3 /flagos-workspace/scripts/install_component.py --component flagtree --action verify --json"

步骤 2 — 停止服务（如在运行）

bash

docker exec $CONTAINER bash -c "pkill -f 'vllm\|sglang\|flagscale' 2>/dev/null; sleep 3"

步骤 3 — 执行安装/升级

FlagGems 安装

bash

docker exec $CONTAINER bash -c "PATH=/opt/conda/bin:\$PATH python3 /flagos-workspace/scripts/install_component.py \
    --component flaggems --action install --json"

指定版本：

bash

docker exec $CONTAINER bash -c "PATH=/opt/conda/bin:\$PATH python3 /flagos-workspace/scripts/install_component.py \
    --component flaggems --action install --version 4.2.1rc0 --json"

FlagGems 升级

bash

docker exec $CONTAINER bash -c "PATH=/opt/conda/bin:\$PATH python3 /flagos-workspace/scripts/install_component.py \
    --component flaggems --action upgrade --json"

FlagTree 安装

bash

docker exec $CONTAINER bash -c "PATH=/opt/conda/bin:\$PATH python3 /flagos-workspace/scripts/install_component.py \
    --component flagtree --action install --vendor nvidia --json"

--vendor 根据 gpu.vendor 自动选择，支持：nvidia, iluvatar, mthreads, metax, ascend, tsingmicro, hcu, enflame, sunrise, amd, xpu。

源码编译（无预编译包时）：

bash

docker exec $CONTAINER bash -c "PATH=/opt/conda/bin:\$PATH python3 /flagos-workspace/scripts/install_component.py \
    --component flagtree --action install --vendor ascend --source --json"

FlagTree 卸载

bash

docker exec $CONTAINER bash -c "PATH=/opt/conda/bin:\$PATH python3 /flagos-workspace/scripts/install_component.py \
    --component flagtree --action uninstall --json"

卸载时自动恢复原版 triton（从备份恢复）。

FlagTree 验证

bash

docker exec $CONTAINER bash -c "PATH=/opt/conda/bin:\$PATH python3 /flagos-workspace/scripts/install_component.py \
    --component flagtree --action verify --json"

步骤 4 — 验证安装结果

FlagGems 验证：

bash

docker exec $CONTAINER bash -c "PATH=/opt/conda/bin:\$PATH python3 /flagos-workspace/scripts/install_component.py \
    --component flaggems --action verify --json"

输出包含版本和 API 兼容性信息（has_enable, has_only_enable, enable_params）。

步骤 5 — 写入 context.yaml

从 JSON 输出中提取字段，写入 component_install、flagtree、environment 等字段。

完成条件

目标组件已确定
安装/升级操作已完成
版本变化已记录
API 兼容性已检查（FlagGems）
context.yaml 已更新
对应 trace 文件已写入

故障排查

问题	解决方案
pip install flag-gems 失败	检查网络连通性和 pip 源配置
FlagTree 安装后 import triton 失败	`install_flagtree.sh verify` 检查，`uninstall` 恢复原版
FlagTree 无预编译包	使用 `--source` 源码编译
版本冲突	先 `pip uninstall` 再重新安装

Security

Security scanPassed, 90 of 100Pattern checks on the skill's files.
Model reviewPassed

See the full security report

Quality evaluation

Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

See evaluation results

Source

SKILL.md in the repositoryCheckingskills/flagos-component-install/SKILL.md

Popularity

7-day trend score: 0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

Claude Code
Cursor
GitHub Copilot
Windsurf
Codex
Gemini CLI
Cline
Amp
Roo Code
Goose
Continue
Kilo Code
Junie
Trae
OpenHands
Qwen Code
AI Agent
Tabnine Chat
Sourcegraph Cody
Amazon Q
Firebase Genkit
Sweep
Devin
Phind
Mentat
Claude Desktop
Aider
GPT Pilot
Plandex
bolt.new
v0
Replit Agent
Anse
TypingMind
BetterChatGPT
LibreChat
Warp AI
Pieces
Fig AI

Embed badge

[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/flagos-ai/flagrelease/flagos-component-install/badge)](https://verified-skill.com/skills/flagos-ai/flagrelease/flagos-component-install)

Something wrong with this skill or its page? Tell us.