Agent, workflow, and developer templates¶
MOI includes built-in starter scenarios in the conversation home screen, preconfigured system agent setups, workflow templates, and developer integration starters so teams can adopt working patterns without starting from a blank page.
1. Agent conversation starter templates¶
On the New conversation page, MOI provides five starter scenario groups (preloaded with sample files and editable prompt templates):
Category |
Built-in starter scenarios |
Typical output |
Reference |
|---|---|---|---|
Document processing |
|
Structured summary tables, issue lists, and source citations |
|
Agent builder |
|
Drafts an agent specification from conversation and saves it directly |
|
Data analysis |
|
SQL queries, KPI comparisons, and anomaly highlights |
|
Knowledge Q&A |
|
Grounded answers with citations and topic breakdowns |
|
Workflow & engineering |
|
Tool invocations, issue/lead drafts, and external callbacks |
Availability depends on the sample assets enabled in your workspace deployment. When a card is not displayed, upload your own sample files and adapt the corresponding prompt structure.
2. Built-in system agent setups¶
MOI ships three system agent setups for engineering operations and repository collaboration (active once required channel slots are bound):
System agent |
Purpose |
Required channel / tool slots |
Trigger mode |
|---|---|---|---|
GitHub Issue Operator |
Triage, organize, create, and update repository issues |
GitHub App / token channel + repository write permissions + callback rules |
Webhook / callback events |
GitHub PR Reviewer |
Inspect pull request diffs and summarize review findings |
GitHub repository channel + code analysis capability |
Pull request events |
Ops Agent |
Patrol suite deployments across Kubernetes and Grafana logs |
Read-only Grafana + read-only Kubernetes + read-only GitHub + WeCom notification channels |
Scheduled patrol ( |
See Tools and Automation for binding and trigger setup.
3. Workflow system templates (10 built-in flows)¶
Focus |
Template pages |
Summary |
|---|---|---|
Text & multimodal RAG |
Document Knowledge Base Preparation, Document Knowledge Base Preparation (Image Index) |
Parse, chunk, vectorize, and index documents and embedded figures |
Structured extraction |
Document Information Extraction, Extraction to Table, Multimodal Extraction |
Extract schema fields to JSON files or Catalog tables |
Drawings & media |
Parse engineering drawings or transcribe audio/video into knowledge bases |
|
Dataset & memory curation |
Generate fine-tuning/eval pairs or curate persistent session memory |
4. Developer and CLI integration starters¶
Python & Go SDK starters: Install the verified offline packages and run workspace, catalog, workflow, and agent examples in AI Studio SDK quickstart and AI Studio SDK guide.
MOI-CLI automation starters: Inspect workspaces, upload data, and trigger runs from shell scripts in CLI quickstart and Complete your first workspace check with MOI-CLI.
Genesis OpenAI / Anthropic compatibility starters: Connect external tools and IDEs via Genesis quickstart, Codex App, and WorkBuddy.