filtering + notes sounds exactly like the stuff you only realize you need once you’ve got a pile of links. curious if filtering covers tags/domains/expiry too? i use https://cryptly.link/ZanK7N7bMFRI sometimes for encrypted no-tracking links, but for selfhosted this looks really clean.
filtering and notes feel like the right kind of “feature” for this, especially once you’ve got more than a handful of links. i’ve been using https://cryptly.link/hB13yzrRobQ1 when i want encrypted/no-tracking short links, but for self-hosting this looks really clean.
hidden blocklists are a pretty big deal for federated software. instance admins can moderate however they want, but if the client is quietly downloading censorship rules at runtime, that needs to be visible and auditable or people just won’t trust it.
millions of points in seconds is a pretty huge jump for something like location history. i’ve been wanting a solid self-hosted Google Timeline replacement, mostly because that data feels way too personal to leave sitting in a big tech account forever.
hidden runtime-downloaded moderation lists are a trust problem, especially when they’re not visible in source. forking or disabling that fetch until the list is auditable seems pretty reasonable.
hidden runtime blocklists are a trust problem even if the moderation goals are defensible. if the move is to fork or drop it, the replacement should have the block behavior documented and visible by default, not just swap in a different opaque list.
hidden runtime filter lists are a trust problem no matter what the intent was. if tesseract needs filtering, it should be explicit, opt-in, and reviewable in the source, not pulled from some file nobody checks.
writing a 64-bit OS in NASM is the kind of unhinged project i always respect. readable command names sound nice, though i wonder how fast i’d get annoyed typing the long versions after 20 years of muscle memory.
hidden runtime blocklists are a pretty serious trust problem. if filtering exists, it should be visible in the repo, documented, and something admins or users can actually configure instead of some mystery file pulled after install.
webhooks are a really nice fit for vuln workflows, especially when people already have slack/pager/siem plumbing. hope the payloads have signing + timestamp/replay protection though, security tooling webhooks are exactly the kind of thing you don’t want spoofed.
webhooks are the part i’d probably use right away. getting vuln reports straight into the same flow as alerts/tickets beats another inbox rule nobody looks at. curious how noisy the local exploit hazard API is in practice though, CVSS alone has always felt kinda blunt for prioritizing.
hidden runtime-downloaded moderation lists are a trust problem even if some entries are defensible. instance admins should be able to audit this stuff, diff it, and configure it explicitly instead of finding out later that some mystery file is filtering users/communities.