Community Continuity

The RLSH community's own infrastructure decays and gets squatted while nobody is watching, and that is fixable at zero users — nobody has to install anything for a link to point somewhere safe instead of somewhere taken over.

The list itself is on the about page, not here. This page is the doctrine behind it. The reader who needs the links is not the reader who needs this.

The finding, checked rather than asserted

While writing this: therlsh.forumotion.com redirects to therlsh.board-directory.net — a squatter, not a continuation of the forum. The last good copy of the real site predates the takeover by about twenty months. Separately, superheroesanonymous.org — the domain a reader would guess for an organisation that is alive and well at .com — has no DNS record at all, which makes it available to anyone who wants that name's credibility tomorrow.

Neither of those is a hypothetical. Both were found by trying the links.

Why it is data, not a page of prose

The first version of this was a markdown table with a hand-typed "checked on" beside each row: a page about link rot, which rots. It shipped three defects that re-reading the sentences could never have found — a live entry with no link at all, two archive links left on plaintext http://, and a liveness note that read as doubt when the evidence was actually strong.

The list now lives in web/src/lib/community.ts as typed data, and community.test.ts asserts against the built HTML what prose cannot:

• no page on this site links a domain on the blocklist, live or dead
• every URL is https
• every live entry is actually rendered somewhere a reader can reach — an entry present in the data and absent from the page is not published, which is verification.md's rule applied to a link list
• no check has been left to go stale in silence. Past six months the build fails until a person re-verifies the links. That rule fails on its own with no code change, which is the intent: a squatted domain sitting unnoticed for years is the exact failure this exists to prevent, and a comment asking someone to re-check does not prevent it

Three rules

Links, never mirrors. A live property gets a link and nothing else. No copy, no scrape, no re-host. propagation.md §3 already sets this for Herocore — NavCom captures the patrol and hands the record to where the community already gathers — and it extends to every entry.

Archive links, not archived content. Where a property is gone, the destination is the Internet Archive's copy, which already exists and belongs to nobody here. Re-hosting a dead forum's actual content on NavCom's own infrastructure is a different act: archiving something abandoned needs nobody's permission, but republishing it wholesale waits on finding an original maintainer and asking. That is real, separate, human work and it is not attempted.

No claim about who is real. Being listed is not a credential and absence is not a judgement — the same refusal positioning.md makes and the same one attestation.md derives everywhere else. This says where things are, not who counts.

How liveness is recorded

Three markers, because collapsing them loses the distinction that got this wrong once:

The middle one matters. Reporting a Cloudflare challenge as "could not confirm" understates what is actually known, and understating is as inaccurate as overstating — invariant 9's distinction between stale and unknown, aimed at a link instead of a shelter. This project has already been bitten by the same confusion from the other direction; see the docblock in web/src/lib/version.test.ts.

What this does not do

• No re-hosted copy of any dead forum's content — gated on maintainer consent, above
• No monitoring. The staleness rule forces a periodic human re-check; it does not watch the domains. Nothing here notices a takeover the day it happens
• No coverage of gear, legal reference, training or OPSEC content. Adjacent RLSH research proposes all four. They are separate, larger pieces of work and none of them is implied by this page existing