Fabro Workflow Factory

What it does

Fabro provides a CLI and runtime to define and run AI-driven workflows described as Graphviz DOT graphs. Workflows can include branching, loops, human-in-the-loop approval gates, model routing via stylesheet rules, sandboxed execution in cloud VMs, and automated retrospectives. It turns LLM turns into coordinated pipelines so agents execute complex tasks while humans intervene only when necessary.

When to use it

Use Fabro when you need structured, repeatable AI workflows that combine planning, implementation, review, and testing stages; when you want human approval gates for safety; or when you need to run agent steps in isolated sandboxes for reproducibility and security. It's ideal for multi-model orchestration, automated test/fix loops, and long-running developer workflows.

What's included

• Scripts: none detected in this package (has_scripts=false)
• References: none included (has_references=false)
• Instructions: installation steps for Claude/Codex/Bash, CLI command examples (run, runs, approve, ssh, retro), Graphviz DOT patterns, model_stylesheet usage, human gate handling, sandbox config, and troubleshooting tips.

Compatible agents

Likely compatible with agent runtimes and CLI-driven LLM integrations such as Claude Code, Codex, and other LLM CLIs that can run shell installers or accept scripted prompts.