Your coding agent shouldn’t write AI workflow plumbing. It should write the workflow.

Описание и применение

Разработчики могут описывать рабочие процессы на естественном языке, а агент Loom генерирует читаемый скрипт, который компилируется в граф агентов, устраняя необходимость писать plumbing‑код и упрощая проверки безопасности и утверждения.

Задача

Developers struggle to build and review AI workflow graphs, spending time translating high‑level descriptions into plumbing code and answering framework‑specific questions.

Результат

Developers can define workflows in plain English, generate a concise Loom script, and rely on Loom to compile and run the graph, reducing plumbing code, improving safety, and streamlining approvals.

Рабочий процесс

• 1. Describe the problem in English to the coding agent. 2. Agent writes a Loom workflow script. 3. VS Code extension validates syntax and visualises the agent graph. 4. Loom compiles the script into executable agents. 5. Agents execute the workflow, performing diagnostics, research, script authoring, safety checks, and human approvals.
• 6. If approved, the script is saved and optionally run.
• 7. Results are reported back to the user.

Доказательства

What if the coding agent didn’t write the implementation at all? What if it wrote the workflow itself, in a language you can read in a minute and see as a graph, and the framework turned that into a running system?

Источник

The VS Code extension validates it as you type and draws the network of agents beside the editor

Источник

budget { tokens: 400000 calls: 60 }

Источник

tool Inspect { use: shell allow: "df, free, uptime, ..." }

Источник

guard { pii: mask }

Источник