Skip to content
View Santhosh595's full-sized avatar

Highlights

  • Pro

Block or report Santhosh595

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Content in all repositories owned by your account will be closed.
Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Santhosh595/README.md

Hi, I'm Santhosh πŸ‘‹

I work at the intersection of cybersecurity and AI β€” building tools that help defenders move faster: autonomous SOC agents, offensive/defensive security tooling, and AI systems that make security operations less manual.

πŸ”­ What I'm working on

  • AI for security operations β€” an autonomous agent that investigates and triages Splunk alerts end-to-end, cutting first-pass triage from ~45 minutes to under 2
  • Python security tooling β€” a modular toolkit spanning file encryption, network intrusion detection, port scanning, and a unified control-panel dashboard
  • Multi-agent RL β€” training AI agents to make routing and logistics decisions in dynamic, disruption-prone supply chain environments (OpenEnv v2.0)
  • Supply-chain security for AI coding tools β€” building hook-scanner, a static audit CLI that inspects what AI agents auto-execute (Claude Code hooks, VS Code tasks, npm install scripts, GitHub Actions) for download-and-execute patterns before they run

🧰 Tech I build with

Python Β· Flask Β· PyTorch (reinforcement learning) Β· Scapy Β· Splunk MCP Server Β· REST APIs Β· Linux Β· Git

πŸš€ Featured projects

Project What it does
SOC-AutoPilot Autonomous AI security agent that triages Splunk alerts in under 2 minutes β€” uses the Splunk MCP Server, a hosted Gemini model, and a self-learning knowledge base to investigate, classify, and generate detection rules
SecureNET-Toolkit- Python-powered security suite: AES-256 file encryption (FileGuard), real-time network intrusion detection (NetworkSniffer), port scanners, security header analyzers, and a unified control-panel dashboard
THREATLENS Flask web app that analyzes URLs for phishing indicators (missing HTTPS, shortened links, suspicious keywords) and returns explainable, risk-based recommendations
Supply-Chain-Disruption-Management-OpenEnv-v2.0 Pre-trained REINFORCE policy checkpoints for the OpenEnv v2.0 supply-chain disruption environment
openenv-project The full multi-agent RL environment behind the v2.0 checkpoints β€” 8-action decision space, 6-type disruption engine, REINFORCE training code, FastAPI server, and HF Spaces deployment
hackerRack-support-triage Evaluator-ready support triage system for HackerRank, Claude, and Visa domains β€” a deterministic, safety-gated pipeline that escalates high-risk or ambiguous tickets instead of guessing
verix Multi-modal evidence review for insurance damage claims β€” adjudicates car, laptop, and package claims from photographic evidence with a transparent reasoning trail (HackerRank Orchestrate hackathon)
hook-scanner Stdlib-only static audit CLI for AI-codetool supply-chain risk β€” scans what agents trust and run: Claude Code hooks, VS Code tasks/extensions, npm install scripts, GitHub Actions workflows, and agent instruction files

πŸ“« Find me

Pinned Loading

  1. hackerRack-support-triage hackerRack-support-triage Public

    Support Triage Agent: safety-gated, deterministic triage of support requests β€” prompt-injection detection, domain/request-type classification, and escalation-or-respond routing.

    Python

  2. SecureNET-Toolkit- SecureNET-Toolkit- Public

    python-powered security suite featuring FileGuard and NetworkSniffer.

    Python

  3. SOC-AutoPilot SOC-AutoPilot Public

    SOC AutoPilot β€” Autonomous AI security agent that reduces Splunk alert triage from 45 minutes to under 2 minutes. Uses Splunk MCP Server, Foundation-sec hosted model, and a self-learning knowledge …

    Python

  4. THREATLENS THREATLENS Public

    A Python Flask web app that analyzes URLs, detects phishing indicators, and provides risk-based recommendations.

    Python

  5. verix verix Public

    Evidence-first multi-modal claim adjudication for insurance damage (car/laptop/package) β€” visual evidence outranks claim text, with prompt-injection detection and explainable structured decisions.

    Python