batch A audit: canonical entry and proof-route repositories

[Riverbraid]

Status

BATCH A AUDIT INVENTORY ITEM / PARTIALLY REMEDIATED

Scope

Batch A covers:

• Riverbraid-Core
• Riverbraid-Evaluation-Kit
• Riverbraid-Verification-Suite
• Riverbraid-Documentation
• Riverbraid-Golds

Confirmed public entry strengths

• The organization profile routes first-time readers to Evaluation Kit, Documentation, Core, Golds, Safety Example, and Gold V2.
• Evaluation Kit is presented as the public entry and reproduction path.
• Documentation points readers to Evaluation Kit and Phase 4 scaffold surfaces, including remediation and review-pass addenda.
• Core states its authority boundary and non-claims.
• Verification Suite and Golds link back to Evaluation Kit or Documentation.

Current finding state

1. Evaluation Kit registry is a pinned snapshot and does not include the later Refusal Gold fail closed patch. Still tracked separately as Riverbraid-Evaluation-Kit#10.
2. Evaluation Kit registry verification commands have uneven verification strength. Classified but not upgraded in Riverbraid-Evaluation-Kit#11 and docs/VERIFICATION_DEPTH_CLASSIFICATION.md.
3. Core workflow clones sibling repositories from mutable default branch state. Still tracked as Riverbraid-Core#6.
4. Evaluation Kit Docker base image remains tag pinned and environment lock marks digest as UNPINNED. Still tracked as Riverbraid-Evaluation-Kit#9.
5. Evaluation Kit verifier dispatch was patched to use a bounded command allowlist. Still tracked as Riverbraid-Evaluation-Kit#8 because execution evidence is required.
6. Verification Suite workflow and target were remediated with a bounded scaffold check and secretless workflow path, but execution evidence is still required.
7. Riverbraid-Golds no longer prints External Alignment: Verified (Simulated); it now prints External Alignment: SIMULATED_NOT_EVIDENCE.
8. Central community health coverage now exists in .github for SECURITY.md, CONTRIBUTING.md, and SUPPORT.md. Per-repo root coverage and root LICENSE coverage remain policy decisions and audit items.

Search-limited non-findings

Search did not return visible hits in Batch A for common private key markers, GitHub token markers, AWS secret marker, npm token marker, pull_request_target, permissions: write-all, secrets.GITHUB_TOKEN, or pipe-to-shell patterns.

This is search-limited and does not prove absence across history, settings, artifacts, dependencies, or unindexed surfaces.

Required follow-up

• Run execution evidence for patched workflow and verifier paths.
• Keep registry freshness locked until an explicit gate exists.
• Verify Refusal Gold patched behavior before any registry refresh.
• Resolve Docker digest evidence before Dockerfile pinning.
• Decide central-only versus per-repo community health and license surfaces.
• Populate readiness matrix with evidence.

Boundary

This issue records audit inventory only.
It does not claim Batch A is secure, complete, production ready, externally audited, or free of defects.
It does not mutate registry, verifier behavior, protocol, hash, seal, manifest, tag, or release state.