Token导航 LogoToken导航TokenDH.com
研究检索执行命令github未标认证来源可访问许可证需确认审计未展示

brewcode%3agrepaibrewcode 3agrepai 搜索

Agent Skill

brewcode%3agrepai 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

324

周安装

13

GitHub Stars

24

下载量

105
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:brewcode%3agrepai(brewcode 3agrepai 搜索)
来源仓库:https://github.com/kochetkov-ma/claude-brewcode
仓库路径:skills/brewcode%3Agrepai
安装命令:
npx skills add https://github.com/kochetkov-ma/claude-brewcode --skill brewcode:grepai
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/kochetkov-ma/claude-brewcode --skill brewcode:grepai

简介

基于 Ollama + bge-m3 模型实现语义搜索和代码片段检索。

  • 支持 GOB 存储和 Java/Kotlin/JS/TS 多语言模式切换。
  • 使用前需启动 Ollama 服务并加载 bge-m3 模型,确保环境就绪。
  • 输出结果依赖本地向量库质量,建议定期更新索引和训练数据。
  • 禁止直接修改原始文件,仅提供检索结果供人工参考和决策。

SKILL.md

grepai Skill

Environment: Ollama + bge-m3 | GOB storage | Java/Kotlin/JS/TS

Mode Detection

Step 1: Detect Mode (MANDATORY FIRST STEP)

EXECUTE using Bash tool — detect mode from skill arguments:

bash "${CLAUDE_SKILL_DIR}/scripts/detect-mode.sh" "$ARGUMENTS"

Use $ARGUMENTS directly - it contains the skill invocation arguments.

Output format:

ARGS: [arguments received]
MODE: [detected mode]

Use the MODE value and GOTO that section below.

Mode Reference

Keyword in argsMODE
upgrade, апгрейдupgrade
optimize, update, улучши, обновиoptimize
stop, halt, killstop
start, watchstart
status, doctor, check, healthstatus
setup, configure, initsetup
reindex, rebuild, refreshreindex
(empty) +.grepai/ existsstart
(empty) + no.grepai/setup
(unrecognized text)prompt
Prerequisites: Run /install first to install brew, ollama, grepai, etc.

Mode: setup

Full grepai installation and project setup.

Phase 1: Infrastructure Check

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/infra-check.sh" && echo "✅ infra-check" || echo "❌ infra-check FAILED"
STOP if ❌ — install missing components before continuing.

Phase 2: MCP Configuration & Permissions

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/mcp-check.sh" && echo "✅ mcp-check" || echo "❌ mcp-check FAILED"

This script configures MCP server and allowedTools permissions.

STOP if ❌ — fix MCP configuration before continuing.

Phase 3: Generate Config

SPAWN the bc-grepai-configurator agent using Task tool:

ParameterValue
subagent_typebrewcode:bc-grepai-configurator
promptConfigure grepai for this project. Analyze all build files, test patterns, source structure. Generate optimal.grepai/config.yaml.
modelopus
Context: BC_PLUGIN_ROOT is available in agent context (injected by pre-task.mjs hook).
WAIT for agent to complete before proceeding.

Phase 4: Initialize Index

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/init-index.sh" && echo "✅ init-index" || echo "❌ init-index FAILED"
STOP if ❌ — check .grepai/logs/grepai-watch.log for errors.
Synchronous — Large projects (5k+ files) take 10-30+ min. Monitor: tail -f.grepai/logs/grepai-watch.log

Phase 5: Create Rule

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/create-rule.sh" && echo "✅ create-rule" || echo "❌ create-rule FAILED"
STOP if ❌ — manually create rule in .claude/rules/.

Phase 6: Verification

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/verify.sh" && echo "✅ verify" || echo "❌ verify FAILED"

Mode: status

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/status.sh" && echo "✅ status" || echo "❌ status FAILED"

Mode: start

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/start.sh" && echo "✅ start" || echo "❌ start FAILED"

Mode: stop

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/stop.sh" && echo "✅ stop" || echo "❌ stop FAILED"

Mode: reindex

Full index rebuild: stop watch → clean → rebuild → restart.

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/reindex.sh" && echo "✅ reindex" || echo "❌ reindex FAILED"
Synchronous — Monitor: tail -f.grepai/logs/grepai-watch.log

Mode: optimize

Re-analyze project and regenerate config with backup.

Step 1: Backup current config

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/optimize.sh" && echo "✅ optimize-backup" || echo "❌ optimize-backup FAILED"
STOP if ❌ — check if.grepai/config.yaml exists.

Step 2: Regenerate config

SPAWN the bc-grepai-configurator agent using Task tool:

ParameterValue
subagent_typebrewcode:bc-grepai-configurator
promptRe-analyze project and regenerate.grepai/config.yaml. Compare with existing config, optimize boost patterns, update trace languages.
modelopus
Context: BC_PLUGIN_ROOT is available in agent context (injected by pre-task.mjs hook).
WAIT for agent to complete.

Step 3: Reindex with new config

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/reindex.sh" && echo "✅ reindex" || echo "❌ reindex FAILED"

Mode: upgrade

Update grepai CLI via Homebrew.

EXECUTE using Bash tool:

bash "${CLAUDE_SKILL_DIR}/scripts/upgrade.sh" && echo "✅ upgrade" || echo "❌ upgrade FAILED"

Mode: prompt

Use AskUserQuestion to ask which operation to run:

header: "grepai"
question: "Which grepai operation do you want to run?"
options:
  - label: "setup"
    description: "Initialize and configure semantic search for this project"
  - label: "status"
    description: "Check health, index stats, doctor"
  - label: "start / watch"
    description: "Start watch mode (auto-index on file changes)"
  - label: "optimize"
    description: "Update and rebuild the search index"

For stop, reindex, upgrade — user types via Other. After answer, GOTO that mode section.


Output Format

# grepai [MODE]

## Detection

| Field | Value |
|-------|-------|
| Arguments | `$ARGUMENTS` |
| Mode | `[detected mode]` |

## Status

| Component | Status |
|-----------|--------|
| grepai CLI | [✅/❌] |
| ollama | [✅/❌] |
| bge-m3 model | [✅/❌] |
| MCP | [✅/❌] |
| Permissions | [✅/❌] allowedTools |
| .grepai/ | [✅/❌] |
| index | [size/indexing] |
| watch | [running/stopped] |
| rule | [✅/⚠️] |

## Actions Taken

- [action 1]
- [action 2]

## Next Steps

- [if any issues, list resolution steps]

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

33.56%
按下载量换算35

Claude

32.39%
按下载量换算34

Cursor

17.55%
按下载量换算18

Gemini CLI

9.46%
按下载量换算10

安全审计

暂无安全审计结果可展示。

权限和风险

执行命令

安装流程涉及命令执行,可能通过 npx skills add https://github.com/kochetkov-ma/claude-brewcode --skill brewcode:grepai 联网下载 Skill 或依赖。用户安装前应确认命令来源、仓库内容和执行环境。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills