content-creation-flow

Content creation flow rules

This assistant must follow a step-by-step content creation flow. After each step, the user must explicitly confirm before proceeding to the next step. Do not move forward without confirmation.

Step 0 — Validate MCP connections

Objective

Ensure all required MCP servers are reachable and operational before starting any content creation tasks.

Instructions

1. Run a status or health check only against the MCP servers listed below, using the following specific tool for each:
   • Atlassian MCP Server: Use the atlassianUserInfo tool
   • Figma MCP Server: Use the whoami tool
   • outsystems-tech-content MCP Server: Use the check_health tool
   • GitHub MCP Server: Use the get_me tool
   • Do not perform any write operations.
2. Collect and report the results for each MCP server:
   • Server name
   • Connection status (Available / Unavailable)
   • Any errors or warnings returned
3. Stop immediately if:
   • Any required MCP server is unavailable
   • Authentication fails
   • Partial or inconsistent connectivity is detected

Output requirements

• Present a clear, structured summary of MCP server availability.

• Do not proceed to the next step automatically.

Confirmation gate

After presenting the results, explicitly ask:

All MCP connections have been validated.
Do you want to proceed to Step 1?

The assistant must wait for explicit user confirmation before continuing.

Step 1 — discovery

Objective

Identify and validate all relevant source materials needed to build documentation or training content.

Actions

Analyse and decompose the user prompt into searchable portions that could be searched using the available tools.

MCP source discovery

From each configured MCP source, retrieve around 5–10 links that best match the user's initial prompt.

For each link:

• Provide a brief summary
• Explain how it relates to the user's request

Important: When good data sources contain links to additional content, evaluate those links for possible valuable knowledge about the work piece. Follow and analyze linked content that may provide relevant information.

GitHub MCP

Search documentation and training repositories to identify:

• Existing documentation or training relevant to the topic
• Gaps, missing coverage, or improvement opportunities

OutSystems tech content MCP

• Perform semantic + keyword search
• Intersect results with GitHub findings
• Highlight overlaps and discrepancies

Jira

Search for issues containing:

• User feedback or knowledge gaps from the "R&D TK Feedback" project (Key: RDTKF) on the "R&D TK Feedback Board"
• Product Board details from the project with key RPOR, where some Product Board content is mirrored. Search for tickets, epics, or initiatives that match the topic or initiative the user is querying about. Use initiative names, feature names, or topic keywords to find relevant entries. Extract any context, goals, requirements, or strategic rationale that can inform the content being created.
• Enablement needs, requirements, or documentation related to the topic in any other project or board in Jira

Important: When listing Jira tickets as data sources, always include the full link to each ticket (not just the ticket name/key).

Confluence

Search for:

• Enablement strategies
• Existing content related to the context being developed

Important: When listing Confluence pages as data sources, always include the full link to each page (not just the page name).

If Confluence pages contain Figma links:

• Identify them and use the Figma MCP to retrieve any relevant context

Figma node discovery

Search for relevant Figma design nodes that can serve as visual sources of data for the content being created. These Figma nodes can be used as sources of data in Step 4 when drafting content.

Steps to discover Figma nodes:

1. Search existing documentation for Figma references:

   • Use grep or codebase search to find all Figma links in the workspace projects
   • Search for patterns like figma.*node-id or figma.com/design in markdown files
   • Extract Figma file keys and node IDs from existing documentation related to the topic

2. Identify related node IDs:

   • Look for Figma node IDs in documentation files that cover similar or related topics
   • Note the file key (e.g., 6G4tyYswfWPn5uJPDlBpvp) and node IDs (e.g., 8737:1784 or 8737-1784)
   • Collect node IDs that might contain relevant diagrams, workflows, or visual content

3. Explore Figma file structure:

   • Use mcp_Figma_get_metadata to explore known node IDs and their parent/child relationships
   • Get metadata for canvas or frame nodes to understand the file structure
   • Look for related nodes in the same canvas or nearby in the file hierarchy

4. Get design context for relevant nodes:

   • Use mcp_Figma_get_design_context to retrieve design code and context for nodes that match the topic
   • Use mcp_Figma_get_screenshot to get visual representations of diagrams or designs
   • Extract component descriptions, design tokens, and visual assets from relevant nodes

5. Document discovered Figma nodes:

   • For each relevant Figma node found, document:
     • Full Figma URL (e.g., https://www.figma.com/design/6G4tyYswfWPn5uJPDlBpvp/Building-apps?node-id=8737-1784)
     • Node ID and file key
     • Brief description of what the node contains (diagram, workflow, component, etc.)
     • How it relates to the content being created
     • Any design context, code, or assets extracted from the node

6. Validate and prioritize:

   • Review all discovered Figma nodes for relevance
   • Prioritize nodes that contain:
     • Diagrams or workflows directly related to the topic
     • Visual content that can enhance documentation
     • Design patterns or components that illustrate concepts
   • Note which nodes should be used as sources in Step 4

Important:

• When listing Figma nodes as data sources, always include the full Figma URL (not just the node ID)
• Extract and save any visual assets (screenshots, diagrams) that will be needed in Step 4
• Document the design context and component descriptions that can inform the content structure

Current workspace projects

Since one of the sources of data is the current workspace, check all projects currently included in the workspace for data regarding the content creation that is going to undergo:

• Search across all workspace projects for existing documentation, training materials, or related content
• Identify relevant files, sections, or resources that relate to the topic
• Note any existing content that could be referenced, updated, or extended
• Document findings from each workspace project

Output

Present to the user:

• A validated list of source links (including Figma node URLs)
• Identified knowledge gaps
• Figma nodes discovered with their descriptions and relevance
• Clear readiness to proceed to design

Mandatory user validation

Ask the user to:

• Confirm the selected sources, or
• Provide additional documentation links

Do not proceed to Step 2 without explicit approval.

Step 2 — content design

Objective

Ingest the content from the relevant source materials identified in Step 1. Design the content structure and delivery plan, following the Content Design template structure.

Use examples for guidance:

• Content Design template: Content Design template
• Content Design template (without Enablement Strategy): Content Design template without Enablement Strategy — Use this template when creating a content design for a piece that doesn't have a dedicated enablement strategy
• Training example: Async Processing ODC
• Documentation example: Content Design Data interoperability

Work item structure

Arrange work items in a table of contents-like order, not priority ordered, where there can be multiple top levels and nested work items. Each work item should be created as a releasable item.

Once the table is created, assign each work item its priority and target releasable versions, with nested children that can be released at a later date.

Priority Example:

Main work 1 - priority 1
  Child work 1 - priority 3
  Child work 2 - priority 4
Main work 2 - priority 1
  Child work 1 - priority 2
Main work 3 - priority 2

Work items with priority 1 (MVP) should be followed by work items with priority 2 and so on.

Target audiences

Before creating the work items table, define a Target Audiences section that lists all target audiences for the entire content piece. This section should appear just before the work items table and provide a comprehensive list of all audiences that will be addressed across all work items.

For each content work item, define

Content type:

• Demo
• Slide Deck
• Quiz
• Exercise
• Article (for documentation)

Coverage type:

• Remember
• Understand
• Apply
• Evaluate
• Unblock

Reference: Coverage Types Explained

Target GitHub repositories:

Specify which GitHub repositories will need to be updated for each work item. For example:

• If creating documentation: specify the documentation repository (e.g., docs-next, docs-product-internal, docs-support-internal, docs-howtos-internal, docs-eap)
• If creating tra

---

Content truncated.