Instruction file imported from ragavsathish/copilot-workflow (.github/instructions/workflow-1-tdd-icv.instructions.md). Copyright stays with the author.

Workflow 1: Technical Design Document + ICV Document

Purpose

Generate two regulated output documents from four input design documents.

Input Documents

Place these DOCX files in workflows/task-1-tdd-icv/input/:

Output Documents

Output 1: Technical Design Document (technical_design.md)

Follow the template structure from extracted/templates/. Map content from inputs as follows:

Output 2: ICV Document (icv_document.md)

Follow the template structure from extracted/templates/. Map content from inputs as follows:

Plan Format for This Workflow

When creating plan.md for this workflow, the generation plan tables must use the mapping tables above. For each output section, specify:

• The exact source file and section number from the extracted input
• A one-line summary of the content that will be written
• Which screenshots (by filename) will be placed in which section

Screenshot Handling

• ICV Steps input will contain screenshots as embedded images
• After extraction, screenshots are in extracted/images/
• In the ICV document, embed each screenshot directly after its corresponding step:

  ### Step 3: Verify Database Connection
  Execute the connection test script and confirm successful output.
  
  ![Database connection verification](images/icv_steps_img_003.png)
  *Figure 3: Database connection test — successful*
  

• Every screenshot must have a figure number and caption

Copilot image tool

Use scripts/image_b64.py to write screenshots into extracted/images/ and create a wrapped base64 text file (<name>.<ext>.b64.txt) plus <name>.<ext>.meta.json. Call example: python scripts/image_b64.py ./icv_steps_img_003.png --out extracted/images --source-input "ICV_Steps.docx#Step 3" --raw Workflow should parse stdout JSON, then use Read to read the .meta.json and the wrapped .b64.txt (concatenate lines to reconstruct continuous base64 for data URIs if needed).

Traceability

At the beginning of each major section, include a traceability note:

<!-- Source: application_design.md, Section 2.1 -->

This ensures auditors can trace every output section back to its source input.

Content Rules

• Use formal, precise language appropriate for regulated documents
• Preserve all technical specifications, version numbers, and configuration values exactly
• Do not invent or assume information not present in the inputs
• If an input is missing or incomplete for a section, write: [To be completed — requires <input document name>]
• Maintain consistent terminology across both output documents