# Agentic Workflows (Slash Commands)
> Automate complex multi-step processes with pre-defined AI recipes.
---
## 🤖 What are Agentic Workflows?
Agentic Workflows are specialized markdown files located in your `.agent/workflows/` directory. They act as **Standard Operating Procedures (SOPs)** for the AI agent. Instead of giving long, repetitive instructions, you can simply trigger a workflow.
### Key Benefits
- **Consistency**: Every release or feature follow the exact same steps.
- **Automation**: Workflows can include shell commands that the AI executes for you.
- **Organization**: Complex checklists are managed separately from your code.
---
## ⚡ How to Trigger a Workflow
You can activate any workflow by using a **Slash Command** in your chat.
For example, if you have `release-prep.md` in `.agent/workflows/`:
1. In the chat, type: `/release-prep`
2. Boring will detect the command, load the file, and start working through the checklist.
---
## 📂 The `release-prep.md` Example
This is a built-in workflow designed to prepare your project for a new version tag.
### What it does:
1. **Version Bump**: Prompts to update `pyproject.toml` and init files.
2. **Doc Sync**: Reminds to update changelogs and README badges.
3. **Verification**: Automatically runs `boring verify` to ensure the release is stable.
4. **Git Ops**: Prepares the staging and tagging commands.
---
## 🛠️ Creating Your Own Workflow
Simply create a new `.md` file in `.agent/workflows/`.
### Template: `my-task.md`
```markdown
---
description: Short description of the command
---
# My Custom Task
Detailed instructions for the AI to follow...
1. [ ] Check file A
2. [ ] Run `npm test`
3. [ ] Report results
```
Now you can run it by typing `/my-task`.
---
## 💡 Best Practices
- **Use Checkboxes**: `[ ]` helps the AI track state across multiple steps.
- **Include Commands**: If a step requires `run_command`, explicitly write it out.
- **Bilingual Support**: If your team is international, you can have `my-task_zh.md` as well.
