docs: document the NeuroSleep ingestion rail and refresh stale counts - #8
Merged
Conversation
The README described neither the NeuroSleep work nor the crates it added, and its headline counts had drifted: it claimed 22 crates, 153 tests and 30 ADRs against an actual 31, 232 and 43. The new entry in "What's built" states the properties that are enforced rather than asserted — only signed derived features are accepted, Helix verifies the signature itself against a separately enrolled key so a bundle carrying its own trust root is rejected, interpretation abstains below the compatible-night floor, and the research visualisation is a physically separate WASM artifact whose entire dependency tree is serde and serde_json. The closing discipline paragraph is extended because "nothing diagnoses" is now checked by a build gate rather than left as a promise, and because the preclinical limit belongs next to the claim it constrains: these features come from an APP/PS1 mouse study, numeric parity is unproven, and that is why the upstream contract is pinned at an exact alpha and the first deployment is an offline research build. The prohibited-claim gate now scans README.md as well. It is the most public surface here, so an overclaim in it would reach more readers than one inside a crate, and it was the obvious omission once the README started describing this rail. Verified by negative control: an injected "Alzheimer risk is elevated" line makes the gate exit 1.
ruvnet
force-pushed
the
docs/neurosleep-readme
branch
from
August 4, 2026 01:45
0528e7b to
f6014b5
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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 README described neither the NeuroSleep work nor the crates it added, and its headline counts had drifted.
Counts corrected: 22 crates / 153 tests / 30 ADRs → 31 / 232 / 43, and
cargo denyadded alongsidecargo auditsince that gate now runs in CI.New "What's built" entry covering the verified ingestion rail, written around what is enforced rather than asserted:
serde+serde_json— no attestation or key-custody code linked.Closing discipline paragraph extended. "Nothing diagnoses" is now a build gate, not a promise, so the README says so — and the preclinical limit sits next to the claim it constrains: APP/PS1 mouse study, unproven numeric parity, exact alpha pin, offline-only first deployment.
Gate gap closed.
check_forbidden_claims.shnow scansREADME.md. It's the most public surface in the repo and was the obvious omission once the README began describing this rail. Verified by negative control — an injected "Alzheimer risk is elevated" line produces exit 1; the scan goes from 7 targets to 8.