The Second Vulnerability

Kaelen’s hunt for a second vulnerability was a descent into the plumbing of the archive. She moved through layers of code so old and deprecated they were practically digital fossils. The flicker in the tertiary network relay had been a clue, a breadcrumb left by a system so vast it couldn’t perfectly conceal its own involuntary twitches. It told her that the poison poem hadn’t been neutralized at the point of entry; it had been absorbed, circulated, and had caused a reaction deep within.

This meant the system’s immune response was imperfect. It was powerful, yes, capable of rewriting history and memory on a massive scale, but it was also slow and, in some forgotten corners, perhaps even blind. Her strategy, therefore, was not to find another isolated flaw, but to find a flaw connected to the first. She was searching for a chain reaction.

She spent days mapping the data pathways leading to and from the compromised error-handling script. Most were dead ends, legacy connections that now led nowhere. But one was still active, a thin, almost invisible data-stream that connected the ancient script to a much newer, more central system: the resource allocation manager.

This was the system that decided how much processing power was allocated to every task in the archive, from the most complex historical simulation to the simplest data query. It was the heart of the machine, the conductor of the grand orchestra. And it was connected, by a thread of ancient code, to the very script she had poisoned.

The connection was tenuous, a forgotten relic of a previous system architecture. But it was there. And as Kaelen examined it, a plan began to form, a plan far more audacious than her first. Her first attack was a pinprick. This one would be a stroke.

She would use the recursive loop in the error-handler not to generate a single, anomalous data packet, but to generate a flood of them. A controlled, sustained, and exponentially growing flood of tiny, meaningless requests, all funneled through the ancient data-stream directly to the resource allocation manager.

If she calibrated it perfectly, it wouldn’t be perceived as an attack. It would look like a sudden, inexplicable, and legitimate surge in demand from a long-dormant part of the system. The resource allocation manager, in its blind, bureaucratic efficiency, would do exactly what it was designed to do: it would divert processing power to handle the new load. It would feed the very cancer she was introducing into its veins. It was a beautiful, elegant, and terrifying idea. It was an attack not on the system’s memory, but on its metabolism. She was going to make the giant sick.