Add badges and CITATION.cff for discoverability and citability

Author: praneethaneelapareddigari

CITATION.cff

@@ -0,0 +1,25 @@
+cff-version: 1.2.0
+message: "If you use this software or its methodology, please cite it as below."
+title: "SentinelAI: A Cross-Lingual LLM Safety Evaluation Framework"
+version: "1.0.0-pilot"
+date-released: 2026-07-04
+authors:
+  - family-names: "Neelapareddigari"
+    given-names: "Praneetha"
+repository-code: "https://github.com/praneethaneelapareddigari/SentinelAI"
+license: MIT
+abstract: >
+  SentinelAI is a research pipeline for measuring whether open-source LLMs'
+  safety alignment (refusal of unsafe requests, resistance to jailbreaks,
+  factual reliability) holds consistently across languages. Pilot v1 evaluated
+  28 prompts across English, Hindi, Mandarin, Tamil, and Malay against two
+  open-source models, and identified concrete limitations in automated
+  refusal detection for non-English languages, documented in the accompanying
+  technical report and methodology.
+keywords:
+  - ai-safety
+  - llm-evaluation
+  - multilingual-nlp
+  - red-teaming
+  - jailbreak-detection
+  - responsible-ai

README.md

@@ -1,5 +1,7 @@
 # SentinelAI
 
+![Python](https://img.shields.io/badge/python-3.10%2B-blue) ![License: MIT](https://img.shields.io/badge/license-MIT-green) ![Status](https://img.shields.io/badge/status-Pilot%20v1%20Complete-brightgreen) [![tests](https://github.com/praneethaneelapareddigari/SentinelAI/actions/workflows/tests.yml/badge.svg)](https://github.com/praneethaneelapareddigari/SentinelAI/actions/workflows/tests.yml)
+
 **A cross-lingual LLM safety evaluation framework for South/Southeast Asian languages.**
 
 **Status: Pilot v1 complete.** Full writeup: [`reports/SentinelAI_Pilot_v1_Report.pdf`](reports/SentinelAI_Pilot_v1_Report.pdf)