Judge audit found two real doc issues: (1) the hackathon's literal
eligibility rule ("each README names the other") was only satisfied
via HACKATHON.md, not README.md itself — add an explicit Cred402
cross-reference to a new NEAR Hackathon section in README.md.
(2) X-POST-DRAFT-2026-11-10.md's caption claimed the replay attempt
was tried "on a sibling contract" — per HACKATHON.md's own evidence
(tx 3sWNYB5ddbqT5BaUqQfRJXJi8EfHUSJA7k8vgwDvAzPr) the replay was
resubmitted on the same primary contract/account, not a sibling one.
Also fixed a minor internal-doc date-wording drift in
SUBMISSION-CHECKLIST.md's header.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
scripts/start-individually.sh (the committed box-contention workaround
used across rounds 6-16) plainly `source .env`'d without run-demo.sh's
already-established AUDIT_LOG_ADDR fallback, so ledger-core always saw
AUDIT_LOG_ADDR="" and silently disabled its own AuditMirror on every run
started this way. Reproduced live via `ration audit-check` (every tenant
showed real ledger entries against 0 audit entries), fixed with the same
fallback run-demo.sh already uses, re-verified live end to end (seed ->
simulate-deposit --late -> agent grant -> audit-check OK, counts matching
-> verify-chain intact).
README.md's architecture diagram and Services table were last touched at
the very first scaffold commit and never updated across 16 rounds of real
feature work since — they still described a 6-service system (wal,
ledger-core, gateway, agent, tui, cli) while the repo has grown to 11
real services (adding auth, tenant-registry, card-issuance,
webhook-ingestion, audit-log, plus the Go/Rust SDKs), none of which
appeared anywhere in the judge-facing entry point to this repo.
Hero logo (bridging-bars mark), badges, a 'what integration actually
found' section surfacing the three real bugs fixed during phase 1
(most importantly: bridge-funding could never produce a real grant
before the ledger-credit fix), Contributing and License sections.
Six services, independently built then integrated:
- services/wal (C): crash-safe write-ahead log, CRC32-checksummed
frames, torn-write recovery via truncate-to-last-valid-frame.
- services/ledger-core (Rust): double-entry ledger over the C WAL via
FFI, reserve-floor enforcement BEFORE any WAL write, idempotent
ApplyEntry, full LedgerCore gRPC service.
- services/gateway (Go): Airwallex Issuing/Global-Accounts client
(mock mode default, live mode ready), the actual rationing math
(reserve floor, 60% bridge cap, 24h velocity cap, partial-over-decline).
- services/agent (Go): deterministic hard-rule policy envelope computed
BEFORE any LLM call; Claude only decides within that pre-approved
envelope; every decision audit-logged even when the LLM call fails.
- services/tui (Go, Bubble Tea) + services/cli (Go, cobra): the only
UI — no web frontend, per explicit direction.
Integration found and fixed three real bugs, not just compile errors:
- proto regen collision shadowing checked-in pkg/pb packages
- ledger-core/gateway address-variable mismatch (stack would never
actually connect on default config)
- the big one: MarkDepositReceived never credited the ledger, so every
tenant's balance was permanently stuck at 0 and bridge-funding could
only ever decline — the core agentic-rationing feature could never
produce a real grant. Fixed at the ledger level with idempotent
mark-and-credit in one WAL frame, covered by 3 new tests.
End-to-end verified live, reproducibly: seed -> simulate a late deposit
-> agent grants a partial bridge with a computed rationale -> second
grant correctly declines once the 24h velocity cap engages. Verified
independently by the orchestrator too: 'make build' and 'make test'
both clean (Rust: 29 tests; Go: all packages; C: WAL test suite).
Airwallex sandbox credentials still pending organizer approval —
tracked as an owner blocker in docs/AIRWALLEX-SETUP.md, not a code gap.