The pulseengine-wasm realm is in the file consumers download - #190
Merged
Merged
Conversation
avrabe
enabled auto-merge (squash)
September 24, 2026 11:12
avrabe
force-pushed
the
feat/wasm-realm-definition
branch
from
September 24, 2026 11:13
e7c63fc to
256081c
Compare
The realm has a published layer, a trust root and a registry, and the canonical `varve-realms.toml` did not mention it. Every document tells a consumer to `gh release download -p varve-realms.toml` rather than paste a key — and that file defined exactly one realm, so pinning the wasm toolchain meant hand-writing realm definitions, which is the thing the file exists to prevent. It also blocks the composition. `compose::walk_installed` resolves an include's realm to a ROOT, and a realm it cannot name is `UnknownRealm`. A layer composing both realms — the worked multi-realm example — cannot be installed by anyone whose realms file knows only one of them. The root is not asserted from the repository that holds it; it is the one that VERIFIED. Layer 2026.09.0 was installed from ghcr.io as a consumer against `f1300acb…` and its signature checked out. That is the only evidence worth having for a trust root. NAMED `pulseengine-wasm`, not `bytecodealliance`, and the comment says why: a realm name is a claim about who vouches, and the key is ours while the payloads are upstream's. The comment also warns that three of its four payloads carry no proof of origin, with their reasons signed into the layer — because a consumer choosing to pin this realm should read them first. `shipped_realms.rs` gates the file we hand out: it parses with the parser that will read it, every realm resolves completely, every registry is an OCI reference, and every root is 64 hex characters. What it deliberately does NOT check is whether a root is the RIGHT one — only the realm holding the private half can say that, and the proof is a layer verifying. Negative-controlled: a truncated root and a registry that is not an OCI reference both fail it. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_019TNtfRjLNhEz82G2ggeeNu
avrabe
force-pushed
the
feat/wasm-realm-definition
branch
from
September 24, 2026 15:29
256081c to
2a53a57
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
The
pulseengine-wasmrealm now has a published layer (2026.09.0, itsfirst), a trust root and a registry — and the canonical
varve-realms.tomldid not mention it.
Every document tells a consumer to
gh release download -p varve-realms.tomlinstead of pasting a key. That file defined exactly one realm, so pinning the
wasm toolchain meant hand-writing realm definitions — the thing the file exists
to prevent.
It also blocks the composition.
compose::walk_installedresolves aninclude's realm to a root; a realm it cannot name is
UnknownRealm. Thecombination layer — one pin over both realms — is uninstallable by anyone whose
realms file knows only one of them.
The root is the one that verified
Not asserted from the repo that holds it. Layer 2026.09.0 was installed from
ghcr.io/pulseengine/wasm-layersas a consumer, againstf1300acb791d330da9b81951d4137f5b1fb0c5efc50213deee4d7a981be28745, and itssignature checked out:
That is the only evidence worth having for a trust root.
Named for who vouches
pulseengine-wasm, notbytecodealliance: a realm name is a claim about whovouches, and the key is ours while the payloads are upstream's. The comment
also warns that three of its four payloads carry no proof of origin, with
the operator's reasons signed into the layer — someone choosing to pin this
realm should read them first.
Gate
shipped_realms.rschecks the file we hand out: it parses with the parser thatwill read it, every realm resolves completely, every registry is an OCI
reference, every root is 64 hex characters. It deliberately does not check
that a root is the right one — only the realm holding the private half can
say that, and the proof is a layer verifying.
Negative-controlled: a truncated root, and a registry that is not an OCI
reference.
fmt · clippy -D warnings · 1035 tests, 0 failed · trace-gate OK.
🤖 Generated with Claude Code
https://claude.ai/code/session_019TNtfRjLNhEz82G2ggeeNu