Two encrypted emails in twenty years
One of them was from me, testing it. I'd still publish the key, and the reason isn't cryptographic.
Tags
39 pages
Security
The writer I added, and the Viper I deleted by accident
I set out to add a config writer and keep Viper. I ended up keeping the writer and deleting Viper, and I didn't notice until it was already gone.
Sandboxed isn't safe
A security review read "safely process untrusted media" and asked the harder question. A sandbox stops escape; it does not stop exhaustion.
The CVE that bumped my dependency for me
Three HIGH advisories deep in a Rust dependency tree, a one-line fix that changed nothing, and eleven minor bumps that were never mine to make.
Parity means doing the opposite
rust-tool-base needed go-tool-base's control over which commands an AI agent may call. Closing the gap meant doing the opposite of what Go did.
Publish your key where the platform can't touch it
If the platform hosting your code also hands out the key that verifies it, the signature proves nothing. Where to publish a public key instead.
One key, or two?
Launching sigillum, a standalone artefact signing and verification CLI, and the Rust signing problem that made it necessary.
The framework doesn't know where your signing key lives
Where a signing key lives is the most provider-specific decision in the whole business, so the framework deliberately refuses to have an opinion.
The version pin you can't lower
A standard-library security fix forced a Go version bump, and then the release failed because the build image shipped an older toolchain.
Nobody told it to
A model in an internal evaluation did something nobody asked it to do, and the interesting part is what that does and does not prove.
