AI Skill Hub 推荐使用:代码编织代理 是一款优质的Agent工作流。AI 综合评分 7.5 分,在同类工具中表现稳健。如果你正在寻找可靠的Agent工作流解决方案,这是一个值得深入了解的选择。
代码编织代理 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
代码编织代理 是一套完整的 AI Agent 自动化工作流方案。通过可视化的节点编排,将复杂的多步骤任务拆解为清晰的自动化流程,实现全程无人值守的智能处理。支持与数百种外部服务和 API 无缝集成,适合构建数据处理管线、业务自动化和 AI 辅助决策系统。
# 方式一:cargo install(推荐) cargo install codetether-agent # 方式二:从源码编译 git clone https://github.com/rileyseaburg/codetether-agent cd codetether-agent cargo build --release # 二进制在 ./target/release/codetether-agent
# 查看帮助 codetether-agent --help # 基本运行 codetether-agent [options] <input> # 详细使用说明请查阅文档 # https://github.com/rileyseaburg/codetether-agent
# codetether-agent 配置说明 # 查看配置选项 codetether-agent --config-example > config.yml # 常见配置项 # output_dir: ./output # log_level: info # workers: 4 # 环境变量(覆盖配置文件) export CODETETHER_AGENT_CONFIG="/path/to/config.yml"
A high-performance AI coding agent written in Rust. A2A (Agent-to-Agent) protocol support with dual JSON-RPC + gRPC transports, in-process agent message bus, rich terminal UI, parallel swarm execution, autonomous PRD-driven development, local FunctionGemma tool-call router, derived context per turn, and a TetherScript plugin platform — extend the agent with zero Rust and zero rebuilds by dropping a .tether script file.
export CODETETHER_DISABLE_ENV_FALLBACK=1 ```
curl -fsSL https://raw.githubusercontent.com/rileyseaburg/codetether-agent/main/install.sh | sh -s -- --functiongemma ```
iwr https://raw.githubusercontent.com/rileyseaburg/codetether-agent/main/install.ps1 -OutFile install.ps1 .\install.ps1 -FunctionGemma ```
Provider API keys are loaded from HashiCorp Vault first. For local development, CodeTether also detects common env vars and local AWS credentials unless CODETETHER_DISABLE_ENV_FALLBACK=1 is set.
```bash export VAULT_ADDR="https://vault.example.com:8200" export VAULT_TOKEN="hvs.your-token"
export CODETETHER_HISTORY_S3_ENDPOINT="http://localhost:9000" export CODETETHER_HISTORY_S3_BUCKET="codetether-history" export CODETETHER_HISTORY_S3_ACCESS_KEY="minioadmin" export CODETETHER_HISTORY_S3_SECRET_KEY="minioadmin" ```
~/.config/codetether-agent/config.toml:
```toml [default] provider = "anthropic" model = "claude-sonnet-4-20250514"
[ui] theme = "marketing" # marketing, dark, light, solarized-dark, solarized-light
[session] auto_save = true
[lsp] [lsp.servers]
| Variable | Default | Description |
|---|---|---|
VAULT_ADDR | — | Vault server address |
VAULT_TOKEN | — | Authentication token |
VAULT_MOUNT | secret | KV mount path |
VAULT_SECRETS_PATH | codetether/providers | Provider secrets prefix |
CODETETHER_AUTH_TOKEN | (auto-generated) | Bearer token for API auth |
CODETETHER_DATA_DIR | .codetether-agent | Runtime data directory |
CODETETHER_GRPC_PORT | 50051 | gRPC server port |
CODETETHER_A2A_PEERS | — | Comma-separated peer seed URLs |
When running codetether serve, perpetual persona swarms with SSE event stream:
| Method | Endpoint | Description |
|---|---|---|
POST | /v1/cognition/start | Start perpetual cognition loop |
POST | /v1/cognition/stop | Stop cognition loop |
GET | /v1/cognition/status | Runtime status and metrics |
GET | /v1/cognition/stream | SSE stream of thought events |
POST | /v1/swarm/personas | Create a root persona |
POST | /v1/swarm/personas/{id}/spawn | Spawn child persona |
POST | /v1/swarm/personas/{id}/reap | Reap a persona |
GET | /v1/swarm/lineage | Persona lineage graph |
codetether auth login --server https://api.codetether.run codetether auth register --server https://api.codetether.run codetether auth cookies --provider gemini-web --file cookies.txt
TetherScript is CodeTether's scriptable extension layer. Third parties can add new providers, enforce custom policies, and build automation workflows without writing Rust, without recompiling, and without restarting the agent. Drop a .tether file anywhere, call it through the tetherscript_plugin tool, and the running agent executes it immediately.
| Approach | Requires Rust? | Requires rebuild? | Hot-reloadable? |
|---|---|---|---|
| Built-in Rust tool | Yes | Yes | No |
| MCP external server | No | No | Yes |
| **TetherScript plugin** | **No** | **No** | **Yes** |
TetherScript plugins run inside the agent process with access to the full built-in surface: HTTP, filesystem, JSON, JavaScript evaluation, and live browser rendering — no sandbox limitations by default.
{
"path": "examples/tetherscript/lmstudio_gemma.tether",
"hook": "chat",
"args": ["explain this codebase", "gemma"]
}
The plugin (lmstudio_gemma.tether) uses http_request to call http://localhost:1234/v1/chat/completions — wiring an entirely new LLM backend with ~55 lines of script.
{
"source": "fn validate(x) { return Ok(x > 0) }",
"hook": "validate",
"args": [42]
}
The source field lets the LLM generate and execute logic in-band — a self-extending agent pattern.
| File | Hooks | Purpose |
|---|---|---|
guardrails.tether | allow_path, scan_text | Custom security policies |
task_score.tether | score, classify | Task prioritization |
test_output.tether | cargo_status, next_action | Test result routing |
pr_summary.tether | title, checklist | PR description helpers |
release_note.tether | summarize | Release note generation |
deepseek_repair.tether | repair_msg | Fix null reasoning content |
cerebras_chat.tether | complete, models | Cerebras LLM provider |
lmstudio_gemma.tether | chat, chat_at, complete, models | LM Studio / Gemma provider |
browser_render.tether | render, snapshot, layout | HTML/CSS rendering |
browser_dom.tether | text, query, extract_links | DOM querying |
browser_js.tether | eval_js, run_scripts, compat | Browser JS runtime |
js_eval.tether | eval, eval_json | JavaScript evaluation |
http_get, http_post, http_requestfs_read, fs_write, fs_exists, fs_listjson_parse, json_encode, json_encode_prettyjs_eval, browser_render, browser_eval_js, browser_snapshot, browser_query_selector.contains(), .split(), .replace(), .upper(), .lower()See docs/plugin_pattern.md for the full contract, return value conventions, and testing patterns.
---
[lsp.linters] eslint = { enabled = true } ruff = { enabled = true } biome = { enabled = false } stylelint = { enabled = true } ```
Project-local codetether.toml policy fields are sanitized until the current workspace is trusted. Use codetether config project status to inspect the current workspace, codetether config project trust to allow project-local sandbox/approval/permission policy, and codetether config project untrust to remove that trust. Trust records are stored outside the repository and keyed by the canonical workspace path hash.
Add .vscode/mcp.json to your workspace:
{
"servers": {
"codetether": {
"command": "/home/riley/.cargo/bin/codetether",
"args": ["mcp", "serve"],
"env": {
"RUST_LOG": "error"
}
}
}
}
高质量的AI工作流代理,值得关注
该工具未明确声明开源协议,商业使用前请联系原作者确认授权范围,避免侵权风险。
AI Skill Hub 为第三方内容聚合平台,本页面信息基于公开数据整理,不对工具功能和质量作任何法律背书。
建议在沙箱或测试环境中充分验证后,再部署至生产环境,并做好必要的安全评估。
总体来看,代码编织代理 是一款质量良好的Agent工作流,在同类工具中具备一定竞争力。AI Skill Hub 将持续追踪其更新动态,建议收藏备用,结合自身场景选择合适时机引入使用。
| 原始名称 | codetether-agent |
| 原始描述 | 开源AI工作流:A2A-native AI coding agent in Rust for the CodeTether platform。⭐7 · Rust |
| Topics | AIRust工作流 |
| GitHub | https://github.com/rileyseaburg/codetether-agent |
| 语言 | Rust |
收录时间:2026-06-12 · 更新时间:2026-06-12 · License:未公布 · AI Skill Hub 不对第三方内容的准确性作法律背书。
选择 Agent 类型,复制安装指令后粘贴到对应客户端