Email tim.almond@thehouseinc.xyz with the details. You will get an acknowledgement within 72 hours and a status update at least every 14 days. Please give us 90 days of coordinated disclosure before publishing; we will credit you in the advisory unless you ask otherwise.

Do not open public issues for suspected vulnerabilities.

Only the latest published release receives security fixes.

Assets. The integrity of erasure-coded data: parity that byte-matches the GF(2^8)/0x11d Reed-Solomon math, and recovery that reconstructs exactly the bytes that were encoded. Consumers (SpaceDB durability, Deputy vault snapshots) stake data durability on both.

Adversaries and untrusted input. This library performs pure computation: it opens no sockets, files, or processes and holds no secrets or key material. The attack surface is its arguments — shard slices, lengths, matrix dimensions, and loss patterns supplied by a caller that may itself be processing hostile data (a corrupted shard store, a malicious peer's stripe).

STRIDE pass.

• Spoofing / Repudiation / Information disclosure: out of scope — no identity, no logging of payloads, no secret state. Shard authenticity is the consumer's job (sign or MAC shards); erasure parity is not a MAC and verify() detects corruption, not forgery.
• Tampering: corrupted shard bytes yield wrong reconstruction only if the caller feeds a wrong-but-well-formed stripe; verify()/pq_check detect inconsistency. Within the library, conformance is gated by byte-identity against ISA-L golden vectors (902-case full grid) so the math itself cannot drift silently.
• Denial of service: hostile dimensions or lengths must never panic or overflow. Every public entry point validates and returns typed errors (CodeError, MatrixError, RecoverError); the no-panic dimension sweep, three fuzz targets (ASan), and Miri gate this. Release builds carry overflow-checks = true.
• Elevation of privilege: the only unsafe code is the SIMD kernels in rusty_erasure-accel, every block SAFETY-commented, every kernel gated byte-identical against the forbid(unsafe_code) scalar core on every architecture in CI.

Residual risks are listed in docs/plans/use-protection-please.md.