Skip to content
View Yeagerist0's full-sized avatar
5 followers · 0 following

Achievements

Achievement: Pull Sharkx2

Achievements

Achievement: Pull Sharkx2

Highlights

  • Pro

Block or report Yeagerist0

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
Yeagerist0/README.md
Neon security console and event graph

HITARTH JAIN

SECURITY ENGINEERING  ×  AI EVALUATION
I build tools that prove their findings — and experiments that show their limits.

Go eBPF AI evaluation Security

Portfolio LinkedIn Kaggle Email

📍 Bengaluru, India  ·  🎓 CS undergrad, class of 2027  ·  🐛 HackerOne + Intigriti


⚡   SELECTED BUILDS   ⚡

🎯 Warrant

An authorization-scoped web pentest agent. Typed probes in, evidence-backed findings out.

Go Safe by scope

🛰️ SentinelX

Self-hosted EDR + lightweight SIEM. Endpoint signals become investigations you can walk.

Go eBPF C

♞ TheKnight

AWS misconfiguration scanner that can ship a Terraform fix as a pull request.

AWS Go Terraform

📡 CyberML

Forecasts near-term attack risk from exposure, honeypots, CVEs and exploit signals.

Python Machine learning

🧪   RESEARCH LAB   🧪

Project The question
Prompt injection in SOC telemetry Can an AI narrator handle hostile endpoint data? 66 payloads, three defenses, and a scoring confound I caught and corrected.
Instruction provenance probe Can GPT-2 tell operator instructions from retrieved text? A length confound led to a documented null result.

🔬 I publish the awkward result too. If the numbers don't support the hypothesis, that's the result.

🌐   OPEN-SOURCE SECURITY   🌐

GitHub Pull Shark achievement, Bronze tier Pull Shark Bronze, level 2 Two published OWASP advisories

21 merged pull requests across security tooling, AI safety, infrastructure and developer tools. Each link goes to the merged change.

View all merged pull requests
Project Merged work
Metasploit Fix Varnish login crash · Remove duplicate Telnet target · Remove duplicate Redis target · Clean up TCP login scanner output
ProjectDiscovery httpx Ignore alternate links as favicons · Avoid retrying absolute favicon URLs as target paths
SentinelX Stabilize Postgres durability CI · Use the Makefile vet target in CI · Reject oversized ingest requests
TheKnight Use the Makefile lint target in CI · Paginate S3 bucket discovery
TensorFlow Validate spaces in eager-mode tf.name_scope
OWASP Agent Memory Guard Bound recursive string conversion · Match immutable keys as globs · Emit security events for swallowed detector errors · Document that snapshot digests are not verified
OWASP Web Security Testing Guide Add SSTI directive-syntax probes · Fix Baidu link-check false positive
OWASP CRS Detect Velocity and FreeMarker directive syntax
NVIDIA garak Load the correct category for SurgeProfanityRacialEthnic
WSTG community fork Fix a typo, reference and list formatting

🧰   TOOLBOX   🧰

Go Python eBPF PyTorch AWS Linux

Also: C/C++ · Rust · Java · TypeScript · Transformers · adversarial testing · reproducible benchmarks

🎲   SIDE QUESTS   🎲

Corrective RAG · JobBoard · Spring Boot expense tracker · Ride-sharing backend · Movie ticket booking

⚡ Build carefully. Measure honestly. Show the evidence. ⚡

Pinned Loading

  1. CyberML CyberML Public

    ML pipeline that forecasts next-week cyberattack risk for public-facing services using OSINT signals — honeypot telemetry, Shodan exposure, CVE feeds, and PoC-exploit tracking.

    Python

  2. JobBoard JobBoard Public

    Job board with role-based employer/candidate flows, built on React + TypeScript + Supabase.

    TypeScript

  3. prompt-injection-soc-telemetry prompt-injection-soc-telemetry Public

    Measuring indirect prompt-injection risk in LLM-narrated security telemetry — a 66-payload corpus, three narrator defense tiers, and an eval harness to compare bypass rates.

    Python

  4. sentinelx sentinelx Public

    Self-hosted EDR + lightweight SIEM that correlates endpoint telemetry into high-signal investigations via a per-host provenance graph. Go + eBPF.

    Go

  5. theknight theknight Public

    AWS misconfiguration scanner that ships the fix as a PR, not just a dashboard.

    Go

  6. instruction-provenance-probe instruction-provenance-probe Public

    Is instruction provenance linearly decodable from GPT-2's residual stream, and is it causal? A length confound found in my own design, then a clean null result.

    Python