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. 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

Santhosh L

Building security tools for the messy edge of AI

GitHub profile Explore projects

About

I build practical tools at the intersection of cybersecurity, AI safety, developer tooling, and automation.

My work focuses on the unglamorous but important parts of modern engineering: detecting prompt injection, tracing AI-agent attack chains, auditing software supply chains, triaging security signals, and turning messy evidence into explainable decisions.

Build small. Verify everything. Ship tools that stay useful.

Featured work

Project What it does
Agentic Attack Chain Correlates multi-step AI-agent attack chains using the ATR Correlation Rule Format.
Hook Scanner Audits AI coding-tool hooks, tasks, extensions, install scripts, Actions, and instruction files for supply-chain risk.
SOC AutoPilot Helps investigate and triage SOC alerts with a human approval gate and a self-learning knowledge base.
CTXProbe Red-team scanner for context poisoning and prompt-injection risks.
THREATLENS Analyzes URLs for phishing indicators and returns explainable, risk-based recommendations.
Verix Evidence-first, multi-modal claim adjudication with prompt-injection detection.
More projects
  • SecureNET Toolkit — standalone Python security utilities for file encryption, network monitoring, TLS/JWT inspection, and security headers.
  • Lean Verify — verifies Lean 4 theorems and definitions with severity-ranked output and CI-friendly exit codes.
  • OpenEnv Project — multi-agent reinforcement-learning environment for supply-chain disruption research.
  • Support Triage — safety-gated, deterministic routing for support requests.
  • Extractor · Extractor 2 — small terminal-first automation tools for monitoring free AI models and digesting technology news.

What I work with

Security       Prompt injection · AI-agent threats · Phishing · SOC automation
Engineering    Python · Flask · GitHub Actions · Linux · CLI tooling
AI systems     LLM agents · MCP · Reinforcement learning · Evidence pipelines
Principles     Human approval gates · Explainability · Reproducibility · Safe defaults

Current focus

  • Making AI-agent behavior easier to observe, correlate, and audit.
  • Building security automation that is deterministic, explainable, and human-controlled.
  • Turning research ideas into small, runnable tools instead of slideware.

Connect

If it touches an AI system, a developer workflow, or a security boundary, I’m probably interested.

Pinned Loading

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

    Support Triage Agent: safety-gated, deterministic triage of support requests — 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 — 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) — visual evidence outranks claim text, with prompt-injection detection and explainable structured decisions.

    Python