taraonsecurity

taraonsecurity

Security researcher writing about web application, cloud, and smart contract security. Notes on vulnerabilities I find interesting and how they get fixed.

Your Releases Ship Weekly. Why Does Your Pentest Run Once a Year?

Picture the usual rhythm. Sometime in the first quarter, someone books a penetration test for June. In June, testers spend two weeks inside your application. In July, a PDF arrives, gets discussed in one meeting, and settles into a shared drive, wh...

Tara reads security 5 minute read 5 Nov 2026 0 comments $0.02

Your Smart Contract Got Audited. Who Tested the Rest of the App?

Application security testing works best as a set of techniques layered across the development cycle. Threat modeling decides what to test. Static analysis, dependency scanning, and secret scanning check the code before it ships. Dynamic, API, and fuz...

Tara reads security 12 minute read 4 Nov 2026 0 comments $0.02

PHP Just Had a Very Bad Week

If you run anything on PHP, you have probably already heard the noise. Two critical flaws in PHP itself, nicknamed PHP2Shell, let an attacker run their own code on your server without logging in and without touching anything more exotic than a web re...

Tara reads security 4 minute read 23 Oct 2026 6 comments $0.54

That "free 1080p" Twitch extension was quietly cc'ing your account to a Russian bot company

Here is a rule of thumb: if a browser extension offers you free 1080p and no ads on a platform that sells 1080p and ads for a living, it wants something back. This time, the something was your account.   Security researchers at Socket fou...

Tara reads security 2 minute read 14 Oct 2026 0 comments $0.04

In Singapore, the Server You Don't Own Is Now Your Problem

For years the easiest way to shrug off a critical system was to point out that it belonged to somebody else. The cloud provider owns the hardware. The vendor runs the platform. The regional system that keeps your Singapore service alive sits in anoth...

Tara reads security 5 minute read 14 Oct 2026 0 comments $0.05

MCP Is Quietly Changing the AI Attack Surface

The Model Context Protocol gives AI agents a standard way to reach real tools and data, which makes MCP security everyone’s problem, including the people nobody told.   Before MCP, connecting an AI assistant to a company’s ticketing system or code re...

Tara reads security 4 minute read 8 Oct 2026 0 comments $0.02

Anthropic Watermarked Claude, Then Told You How to Wash It Out

Every Claude model Anthropic has shipped since 2 August 2026 writes a watermark into its own text. You cannot see it. There is no badge and there are no hidden characters, because the thing lives in the statistical pattern of the word choices themsel...

Tara reads security 2 minute read 30 Sep 2026 0 comments $0.21

IDOR never dies

There is a bug I have been finding for years, and I found it again last month reading the news instead of a client’s app. It is almost embarrassingly simple, and that simplicity is the whole reason it outlives every new framework and every new best p...

Tara reads security 4 minute read 29 Sep 2026 0 comments $0.01