把自然语言描述转为结构化 JSON,并由 mcp-diagram-generator MCP 服务生成 Draw.io、Mermaid 或 Excalidraw 图表文件。
编程
supervisor-config-gen
试用Bash tool that generates a Supervisor (supervisord) `[program:*]` config file from zero CLI flags — it derives everything from the current directory (program name = dirname, `command=<dir>/run.sh`, `directory=<dir>`, `user=$(whoami)`, logs under `$HOME/logs/supervisord/`) and writes `<dirname>-supervisor.conf` into that same directory, overwriting silently if present. Use when the user wants to generate a Supervisor program config for an app directory that has a `run.sh` entrypoint.
它能做什么
supervisor-config-gen — generate a Supervisor program config from a directory
技能文档
supervisor-config-gen — generate a Supervisor program config from a directory
supervisor-config-gen is a single bash script with no flags, no args, no env vars. Run it from inside an app directory and it emits a Supervisor [program:x] config for that app, inferring everything from the directory itself.
Security & safety
- Generates a local plain-text config file. No network calls, no external services.
- The generated config controls what Supervisor executes and as which user — review the output before pointing supervisord at it, same as reviewing any generated shell/config artifact.
- It overwrites
-supervisor.confin the current directory without prompting if one already exists. Don't run it in a directory where that file has manual edits you want to keep. - The script does not sandbox or validate the target
run.sh— it just references its path. Make surerun.shis what you expect before Supervisor runs it.
When to use
- The user wants a Supervisor program config for an app directory that has a
run.shentrypoint (a startup script Supervisor should manage). - The user says "generate a supervisor config for this app" / "supervisorize this" / "make a supervisord conf".
When NOT to use
- The app has no
run.sh— the generatedcommand=will point at a script that doesn't exist. Add/rename arun.shentrypoint first, or hand-write the config. - You need non-default supervisord settings (custom
numprocs, environment vars, non-TERM stop signal, different log paths/rotation, etc.) — this tool has no flags to control those. Generate the base config, then hand-edit it. - You need configs for multiple apps at once — the tool is single-directory, single-invocation.
Inputs → output
There is nothing to pass in. The script derives every field from ambient state at run time:
| Field | Source |
|---|---|
program_name (used in filename + [program:x]) | basename $(pwd) — the current directory's name |
command | $(pwd)/run.sh — a run.sh is REQUIRED to exist in the current directory (the script doesn't check, it just hardcodes the path) |
directory | $(pwd) |
user | $(whoami) — the user invoking the script |
stdout_logfile / stderr_logfile | $HOME/logs/supervisord/-{out,err}.log |
| output path | ./-supervisor.conf (written in the current directory, overwritten if it exists) |
Everything else (numprocs=1, stopsignal=TERM, stopwaitsecs=10, stopasgroup=true, killasgroup=true, autostart=true, autorestart=true, redirect_stderr=true, redirect_stdout=true, 50MB log rotation with 10 backups on stdout) is fixed/hardcoded — not configurable via flags.
Example invocation
Given an app directory shitty-app/ containing main.py and run.sh:
cd shitty-app
supervisor-config-gen
# Generated shitty-app-supervisor.conf
Resulting shitty-app-supervisor.conf (paths reflect the actual invoking user/home/cwd):
[program:shitty-app]
command=/home/not-root/shitty-app/run.sh
directory=/home/not-root/shitty-app
process_name=%(program_name)s
numprocs=1
user=not-root
stopsignal=TERM
stopwaitsecs=10
stopasgroup=true
killasgroup=true
autostart=true
autorestart=true
redirect_stderr=true
redirect_stdout=true
stdout_logfile=/home/not-root/logs/supervisord/%(program_name)s-out.log
stdout_logfile_maxbytes=50MB
stdout_logfile_backups=10
stderr_logfile=/home/not-root/logs/supervisord/%(program_name)s-err.log
stderr_logfile_maxbytes=50MB
For install steps and the full field reference, see references/setup.md.
相关技能
在本地磁盘以分类纯 Markdown 文件保存需要长期留存的事实,与智能体内置记忆并存。
诊断生产力系统反复失效的根因,给出最小干预——容量测算、瓶颈定位、可靠的本地记录。
通过一次 REST API 调用,向 10 个社交平台发布视频、图片、文字与文档。
编写、调试与调优 Playwright 测试,涵盖定位器策略、追踪诊断与 CI 友好的超时配置。
按用户明确指令,在得到大脑(Get笔记)中保存、搜索并管理笔记与知识库。
psyb0t 的更多技能
浏览全部技能对接用户自部署的 mt5-httpapi MetaTrader 5 网关,每次涉及真实资金的写操作都必须逐笔确认后再执行。
面向反爬检测栈 QA 与授权测试场景的 Docker 浏览器自动化工具。
自托管、OpenAI 兼容的语音服务,一个容器搞定转写、翻译与合成。
在固定白名单的 SSH 沙箱里跑 ffmpeg、sox、ImageMagick 处理音视频和图片。
通过 SSH 调用 Qwen3-TTS 生成语音,支持预设音色、声音克隆与声音设计。
一个端点统一管控多个 IMAP/SMTP 邮箱,跨账号并行完成读取、检索、发送、标记与删除。