ABHEDYA

AGENTIC AI & AUTONOMOUS WORKFLOWS:

Agents that plan, use tools, and execute multi-step tasks.

CHANDRAKALA A S

What problem are you solving?

• Security teams receive too many alerts from servers, accounts, networks and cloud systems.

• Manual investigation makes it difficult to connect related events quickly.

• Important attacks can be missed or handled too late.

• Example: repeated failed logins → unusual successful login → sensitive file download.

Alert overload

Many events; limited analyst time.

Slow investigation

Evidence is spread across logs.

Delayed response

Threats may cause more damage.

Describe your solution. How different is it from any of the other existing ideas?

Cyber Sentinel AI uses a coordinated team of AI agents to investigate suspicious activity and support safe response.

1. Log Analysis

Find suspicious patterns

2. Threat Intel

Check IPs / indicators

3. Event Correlation

Connect related events

4. Investigation

Build attack timeline

5. Risk Assessment

Prioritize severity

6. Response

Suggest next actions

Key difference: event correlation + evidence-backed investigation + human-approved response + verification.

List of features offered by the solution

Example alert path

Workflow / Idea

�

Technology stack

• Frontend: React.js + Tailwind CSS

• Backend: Python + Fast API

• Agent orchestration: Lang Graph; LLM: Gemini API

• Detection/data: Scikit-learn + Pandas

• Database: PostgreSQL / Supa base

• Security context: Wazuh + MITRE ATT&CK (integration-dependent)

High-level architecture diagram

Response loop