Hacker

Kalash Poddar

@kalash33

Joined Oct 2026Software Engineer at Texas InstrumentsIndia

Projects

  • Top 50

    Taintgraph

    Graph HacksOct 2026

    Taintgraph shows how a software supply-chain attack reaches your app, and who could poison it next. In September 2025 a phished maintainer pushed malware into chalk, debug and 16 other npm packages; a week later the Shai-Hulud worm began stealing npm tokens and republishing every package its victims maintained. In 2026 the same playbook hit TanStack and LiteLLM. An advisory says a version is bad. It doesn't say how it reaches your app, which maintainer account could do the same to you next, how far a worm would spread, or what exactly to change. Give Taintgraph a lockfile (npm, pnpm, yarn, bun, uv, poetry, pip), a public GitHub repo URL, or an AI tool's MCP config. It builds a FalkorDB graph and four agents work on it: - Scout: ingests and triages. - Analyst: answers questions using Gemini or any LLM over 14 fixed, read-only Cypher tools, showing every step and the query behind it, and checking its answers against the graph. - Reporter: writes the incident report. - Sentinel: re-checks every version ever installed when new advisories appear. You get: - the exact route of each bad version into the app - the accounts and maintainer teams that could poison the tree, plus a wave-by-wave worm simulation - a fix plan with re-checked safe versions and a ready-to-paste overrides block - memory of your verdicts that changes the next scan; malware is never skipped automatically It's for developers and security teams responding to supply-chain incidents, and for coding agents, through its MCP server, that should check a dependency before adding it. Proof: - 19 of 19 real incidents (2018–2026) detected, with 0 false alarms on 10 clean controls. - Scanned real repos (React, Astro, Vite, LangChain, FastAPI, Flask, Bun, TypeScript): only the genuinely compromised lockfiles are called compromised. - 80 automated tests, run in CI.