azure-spec-audit

Extracts text with location traceability from documents and audits them against Azure Specialization Module B controls.

4|1|Updated May 22, 2026
One-click install
npx skills add https://github.com/ThomazRossito/ai-data-agents --skill azure-spec-audit-thomazrossito
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: azure-spec-audit
Source: https://github.com/ThomazRossito/ai-data-agents/tree/main/plugins/ai-data-agents/skills/azure-spec-audit
Command: npx skills add https://github.com/ThomazRossito/ai-data-agents --skill azure-spec-audit-thomazrossito

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires pdfplumber, pymupdf, python-pptx, openpyxl, python-docx, pytesseract, pillow, rapidocr-onnxruntime, and includes scripts (resource) components.

What problem does it solve? Auditing a document dossier against the 7 controls of the Analytics on Microsoft Azure Specialization (Module B) requires manually reading PDFs, Word files, spreadsheets, slides, and screenshots, then mapping each passage to a control with citable evidence. This Skill automates that extraction and audit workflow, producing a coverage matrix with verdicts and evidence citations. ## Core Features & Use Cases - Traceable multi-format extraction: Extracts text from PDF, DOCX, XLSX, PPTX, CSV/TXT/MD and images via OCR, preserving page, sheet, slide, or line location for every excerpt. - Deterministic control verdicts: Maps evidence to the 7 Module B controls using keyword search and a fixed verdict rubric, producing per-control coverage ratings (Complete, Partial, Missing, Needs Review) and unique-customer counts. - Remediation guidance: For every gap, delivers actionable next steps, accepted documentation types, and fillable templates such as SOW, sign-off, SOP, and Assessment Report. - Use Case: A consulting partner preparing for the Azure Analytics Specialization audit drops a folder of customer evidence (screenshots of Key Vault, Data Lake, DevOps pipelines, plus SOWs and design docs) and receives an audit_report.md coverage matrix plus a machine-readable findings.json. ## Quick Start Audit the documents in my input folder against the 7 Module B controls of the Analytics on Microsoft Azure Specialization and generate the coverage report and findings JSON.

Frequently Asked Questions about azure-spec-audit

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I audit documents against the Azure Analytics Specialization Module B controls?▼

Point the Skill at your input directory of evidence documents. It extracts text with page, sheet, or slide traceability, maps excerpts to the 7 Module B controls, and outputs an audit_report.md coverage matrix plus a findings.json with verdicts and cited evidence.

How to extract text from scanned PDFs and screenshots with OCR?▼

The bundled extract.py rasterizes image-only PDFs with PyMuPDF and runs OCR via pytesseract, falling back to rapidocr-onnxruntime when the tesseract binary is unavailable. OCR runs in parallel across files and each excerpt is tagged with its page and an (OCR) provenance marker.

What file formats does the Azure spec audit extraction support?▼

Supported formats are PDF, DOCX, XLSX, XLS, PPTX, CSV, TXT, MD, and images PNG, JPG, JPEG, BMP, and TIFF via OCR. Duplicate files are skipped by content hash, and unsupported extensions are flagged as Needs Review rather than silently ignored.

Does the audit work when evidence covers only one customer?▼

Yes. In single-customer mode the coverage verdict focuses on that customer's dossier, while the unique-customer axis is shown as a known gap (for example 1 of 3 customers) without downgrading the coverage rating.

Why should large PDFs not be read directly during extraction?▼

Reading large binary documents directly can exceed the SDK message buffer and crash the session. The Skill instead extracts via Python into a _work directory, truncates each unit to about 3000 characters, and prints only short progress summaries to stdout.