Test suites, the full-system ceremony, red-team campaigns — proof, honestly stamped · content/evidence.md

Evidence

What we claim, and what proves it. The generated blocks on this page are built from files in the repo at build time and stamped with file dates — nothing here is hand-typed, so it cannot drift from the facts.

Claims → evidence — the receipt for every public claim

Every load-bearing claim on this site, where it is proven, and how you can check it yourself without trusting us. Object/tx ids are real Qalnet devnet records; suite paths are repo files you can run.

One pricing story (Pricing) and the fee table (Fees) are scoped by rows 4–8 above.

Live fleet verdicts

The cross-device use-case harness (messaging, data transfer, NFTs, value transfer, referrals — per device pair across the fleet) writes machine-readable verdicts to testlogs/verdicts.json. When that file exists, the panel below renders it live; until then it says so honestly.

The mint ceremony — one coin through every system

The ceremony dress-rehearses the whole loop on-chain and on the ledger: on-chain issue → cap-attack refusal → top-up → chat transfer → compute escrow → redemption with fees and dust → burn → anchored. This is the proof behind the "cap is enforced in code" claim. (2026-09-15: the ceremony page keeps its historical name — its step-1 on-chain mint predates the redesign; the on-chain twin is being redesigned from mint-gated to genesis-preminted + transfer-gated at the cap-raise checklist.) Latest recorded run:

14/14 steps PASS · evidence file testlogs/last_ceremony.json · last modified 2026-09-12.

Full narrative: Mint Ceremony.

The test suites

Every service ships a proof suite — the "52-test", "58-test", "27-test" figures quoted in the ceremony notes are these suites. The inventory below is generated from the scripts themselves at build time; each one-liner is the script's own header description, and the dates are file mtimes.

39 proof suites ship with the services. One-liners are each script's own header; dates are file mtimes.

Red-team campaigns

Two full audit campaigns so far. Mode for both: findings-only, reversible PoCs, every mutation undone (each file carries an UNDO ledger). Severity counts below are parsed straight from the reports at build time.

Mode for both campaigns: findings-only, reversible PoCs, every mutation undone (UNDO ledgers live in each file). Severity counts are parsed from the reports at build time:

Campaign 2 re-attacked the patched surface: it confirmed the fixes that held and flagged regressions plus new criticals. Open remediation items are tracked in agents/HANDOVER_SYSTEM.md (Track S queue) — this page links to the work, it doesn't spin it.

Open remediation items are tracked in the system track's live queue (agents/HANDOVER_SYSTEM.md) — this page links to the work; it doesn't spin it.

How to reproduce any of this

• bash scripts/status.sh — live probe of every service in the manifest.
• python3 scripts/check_links.py --quiet — verifies every link on every page of this site.
• Any suite in the inventory above: run it (bash <suite>); the evidence page re-reads headers and dates on every build.
• GET :8823/report — the program ledger, to the cent, any time.

Companions: Ground Truth (the numbered fact sheet) · Fees (where money actually moves) · Pricing (the one pricing story) · Paying with Bitcoin (pre-launch policy) · Testing Roadmap & Strategy (what's next to prove).

Proof: this page is the proof hub — the claims→evidence table above, the audit + red-team records below, and the era-1 checkpoint (object 0xe9eff274a90a0fb5d8a52a9fedebd5d5a6bd0880476fe12da70b4a348fab7905, tx 2gopkDw9ecBexPAA3U8vZhnoMqvZ673cKk9278G3NYGS) are the receipts every other page's Proof footer points to. The plan that keeps them true: Testing Roadmap & Strategy.