Sentry (archived)

ARCHIVED. FROZEN. NOT MAINTAINED.

Archived: 26 August 2026 Final version: 0.1.0, wire profile SCIT/1. Never released to Bitcoin mainnet. Final source commit: 4c48ff2 (10 August 2026) Replacement: none. Sentry was retired without a successor.

This repository receives no changes, no fixes, no security patches, and no support. It exists as a permanent public record of a protocol Bitcoin Universe designed and then retired. Nothing here is a live product, a supported library, or a specification you should build against.

Read the security warning before you follow any instruction preserved in this repository.

What Sentry was

Sentry was a Bitcoin protocol designed to sit between a person and a signing decision. The idea: give a user an identity, recorded on Bitcoin, that reads a proposed transaction alongside them, remembers what they have already declared unacceptable, and is structurally incapable of signing anything itself.

The on-chain part was deliberately small. A Sentry had exactly one live control output, called a carrier. Five confirmed Bitcoin transactions covered its whole life:

Event	What it did
BIRTH	Brought a Sentry into existence with provably zero permissions
TRANSFER	Changed its owner, and nothing else
CHECKPOINT	Changed its committed state, with the owner unchanged
FUSE	Retired two to four Sentries and created one broader successor
RETIRE	Ended it and returned the carrier capital in full

Each event was an ordinary Bitcoin transaction carrying one 74-byte OP_RETURN payload inside a 76-byte output script. Bitcoin held a 32-byte commitment to state and never the state itself, so any indexer reading full blocks could rebuild ownership, supply, lineage, and event order without trusting a server.

The wire format was named SCIT/1. The preserved specification is complete enough to write a parser from, and it is still the best description of what Sentry was.

What Sentry was not

This matters more than usual, because the repository was published while the work was still unproven and the published README said so in plain terms.

• There was never a mainnet profile. Not unset, not pending. Absent by design. Activation required a network profile fixing an anchor height, an anchor hash, and a future activation height, and no such profile was ever published for Bitcoin mainnet.
• Nothing was audited. No external review of the design, the reference implementation, or any signing flow ever took place.
• The central claim was never tested. The premise was that a Sentry helps people refuse dangerous approvals without drowning them in false alarms. That is a testable claim about human behaviour. It was never tested.
• There was no signer. The reference implementation held no key, no seed, no wallet call, no broadcaster, and no network client. It read and wrote bytes.
• It was never published to a package registry. See the security warning.

Why it was archived

Sentry did not reach a public release, and the work was retired rather than shipped. On 26 August 2026 the repository contents were removed in commit 7a8c9b5 and the repository was archived. This README and the archive/ directory restore the historically valuable public material so the record is readable without digging through Git history.

The Bitcoin Universe documentation portal publishes no Sentry page. There is no SCIT/1 entry in the protocol atlas, because Sentry never became a protocol the organization ships or indexes.

What replaced it

Nothing. No Bitcoin Universe protocol inherited Sentry's design, its identity model, or its wire format. Statements to the contrary are wrong.

If you arrived here looking for a Bitcoin protocol Bitcoin Universe actually ships, indexes, or documents, the Protocol Atlas lists every one of them and points at the specification that owns each.

Migration guidance

There is nothing to migrate to and nothing to migrate from, because no released software depended on Sentry.

• If you built against SCIT/1: you were building against an unreleased specification with no mainnet profile. Stop. There is no network to index and no supported implementation to track.
• If you have a dependency on @bitcoinuniverse/sentry: remove it. That package was never published. See the security warning.
• If you want a Bitcoin protocol with a real specification, vectors, and an indexer: start at the Protocol Atlas.
• If you want the design ideas: the specification is MIT licensed. Reuse it, fork it, or improve on it. Do not present a fork as a Bitcoin Universe product.

Security warning: read before following anything here

The preserved material was written while the repository was active. Some of its instructions are unsafe or misleading to follow today.

1. Do not run npm install @bitcoinuniverse/sentry. The preserved final README told readers to install that package. It was never published: the npm registry returns HTTP 404 for @bitcoinuniverse/sentry as of 1 September 2026, and the @bitcoinuniverse scope is not held by this organization. Installing that name now would install whatever a stranger later publishes under it. Treat any package by that name as untrusted.
2. Do not run the preserved build and verification commands. The reference implementation, its package.json, its lockfile, and its CI were removed from this repository. npm install && npm run build, npm test, and node cli.mjs verify cannot work here, and archive/examples/quickstart.mjs exits with code 2 because there is no build to load.
3. Do not treat SCIT/1 as a live specification. It defines no mainnet activation, it was never audited, and it will never receive a correction. A defect found in it today will not be fixed.
4. Report nothing here as a vulnerability against a live system. This code runs nowhere. If you find a flaw that also affects a Bitcoin Universe product that is still running, report that product privately using the organization's security policy.

The preserved documents also point at github.com/bitcoinuniverse/.... The organization was later renamed to bitcoinuniverseio, and GitHub still redirects those URLs, so they resolve. The npm scope did not move with it, which is exactly why point 1 matters.

What is preserved here

Everything in archive/ is a byte-identical copy of the file as it stood at commit 4c48ff2, the last commit that carried the protocol material. Checksums are recorded in archive/README.md so any copy can be checked against Git history.

Path	What it is
archive/sentry-protocol.md	The normative SCIT/1 specification, 14 sections
archive/vectors/conformance.json	Conformance fixture: marker script, profile, birth and fusion identities, zero-permission root, BIP 340 cross-check
archive/examples/quickstart.mjs	The lifecycle example, preserved as written and no longer runnable
archive/CHANGELOG.md	The 0.1.0 release record
archive/README-final-2026-08-10.md	The README as published on the final source commit
LICENSE	MIT, restored so the preserved material carries its license

The TypeScript reference implementation, the command line tool, the test suite, and the CI workflows were removed on 26 August 2026 and are not restored here. They remain reachable in this repository's Git history at commit 4c48ff2.

Permanent URLs

These are stable and are expected to keep resolving.

What	URL
This repository	https://github.com/bitcoinuniverseio/sentry
Preserved specification	https://github.com/bitcoinuniverseio/sentry/blob/main/archive/sentry-protocol.md
Final source commit	https://github.com/bitcoinuniverseio/sentry/commit/4c48ff2f0d572ecbee79067eb2239a455c61f085
Content-removal commit	https://github.com/bitcoinuniverseio/sentry/commit/7a8c9b5e42b41d69431616275fe409339828e9ef
Protocol Atlas (what we do ship)	https://docs.bitcoinuniverse.io/protocols/
Documentation home	https://docs.bitcoinuniverse.io
Lifecycle vocabulary	https://docs.bitcoinuniverse.io/status/

License

MIT. See LICENSE. The license applies to the preserved material and survives archival.