| name | requirements-ingest |
| description | Normalize requirements from any format into consistent, chunked representation with traceability and classification. |
| license | MIT |
Requirements Ingest
Transforms requirements documents (PDF/DOCX/Markdown/Email) into structured, atomic chunks with classification and traceability.
Core Function
Input: Raw files + project_id
Primary Output: Markdown format matching original specification (for downstream skills)
Secondary Output: JSON format saved to structured folders (for machine processing)
Output Destination:
- Markdown:
outputs/projects/{project_id}/Analysis/requirements.md (primary for downstream)
- JSON:
outputs/projects/{project_id}/Analysis/requirements.json (machine processing)
Directory Structure: Auto-created project folders with Analysis subfolder containing dual-format requirements, processing logs, and glossary files
Usage
GitHub Copilot Integration (Recommended):
Use this skill directly in Copilot by providing requirements documents.
Copilot will automatically extract and classify requirements using its built-in AI.
Example prompt:
"Use requirements-ingest skill to process this requirements document and return structured JSON with atomic requirements, classifications, and traceability."
Traditional Script Approach:
from requirements_ingest import RequirementsIngestor
ingestor = RequirementsIngestor()
result = ingestor.process_files([], )
