RC RANDOM CHAOS

Articles

Long-form writing on tech, culture, and the edges of the internet.

Your build trusts whatever it can find
supply-chain-securitysystems-drift

Your build trusts whatever it can find

A build system executed code because a name resolved, not because content was verified. How trust attaches to the reference and outlives the artifact.

7 min read
Copilot shipped CWEs in 40% of NYU's 2021 scenarios
AI code reviewsecure coding

Copilot shipped CWEs in 40% of NYU's 2021 scenarios

Why working AI-generated code still gets rejected in security review: functional correctness is not security correctness, and CWEs ride through clean output.

6 min read
CORS misconfiguration is consent, not an exploit
corsweb-security

CORS misconfiguration is consent, not an exploit

CORS misconfiguration explained at the mechanism level: origin reflection, null origin, broken allowlist matching, the credentialed-read exploit path, and why it stays invisible in telemetry.

6 min read
Half of LG TV apps are exit nodes
residential proxiessmart tv security

Half of LG TV apps are exit nodes

Residential proxy SDKs sit in nearly half of LG webOS apps, turning smart TVs into rentable exit nodes LG neither blocks nor detects. The mechanism and telemetry gap.

6 min read
Kees Cook deleted strncpy from the Linux kernel
linux kernelmemory safety

Kees Cook deleted strncpy from the Linux kernel

Linux removed strncpy after six years and 360+ patches. The mechanism: missing NUL termination, out-of-bounds reads, kernel info leaks, and why strscpy replaces it.

6 min read
Sanctioned keylogger, unlocked back end
data exposureinsider risk

Sanctioned keylogger, unlocked back end

Meta's exposed employee keystroke telemetry is not an AI story. It is a third-party data-handling failure: T1056.001 collection, an unauthenticated store, T1530 read.

6 min read
Sandi Metz's 2016 essay is a bug-class taxonomy
heap exploitationmemory corruption

Sandi Metz's 2016 essay is a bug-class taxonomy

The 2016 'wrong abstraction' rule read as a memory-corruption bug class: use-after-free, type confusion, double-free, from DirtyCOW to Dirty Pipe and DirtyCred.

6 min read
The guard checks the badge, never the room
prompt injectionrole confusion

The guard checks the badge, never the room

Prompt injection is role confusion: systems that derive content authority from channel trust execute attacker input as instruction.

7 min read
The log that killed your SSD was working perfectly
systems driftreference-based execution

The log that killed your SSD was working perfectly

A Codex logging bug can write terabytes to a local SSD because the system resolves a configured reference repeatedly and never revalidates the trust behind it.

8 min read
There is no free()
prompt-injectionllm-security

There is no free()

Claude Code's extended thinking is not a use-after-free. The real exposure is indirect prompt injection into a tool-holding agent - OWASP LLM01, ATLAS T0051.

7 min read
They knocked on his door, not his firewall
operational securitythreat modeling

They knocked on his door, not his firewall

Danish police raided privacy activist Lars Andersen with no technical breach. The boundary moved to physical custody, where remote controls were never positioned.

7 min read
This isn't a bug. It's the default.
MITRE ATT&CKEndpoint Denial of Service

This isn't a bug. It's the default.

Codex writes unbounded session logs to local SSDs. Mapped correctly to MITRE T1499, not T1071 - a disk-exhaustion DoS primitive EDR baselines miss.

7 min read