Skip to main content
AVOID.NET
← avoid.net

Verify a decision

Every moderation decision on AVOID.NET is anchored to the Solana blockchain. You don't have to trust us — you can verify cryptographically that we committed to a verdict at a specific moment and have not rewritten it.

How verification works

  1. We commit. When a moderator accepts/rejects a submission, we serialize the decision into deterministic UTF-8 bytes (payload_canonical_string), hash it with SHA-256, encode the digest as base58, and write it to Solana inside an SPL Memo v2 transaction.
  2. We store the bytes. The exact bytes we hashed are stored alongside the decision in our database. Anyone can read them and recompute the hash in any language.
  3. You compare three values. Database hash, your independently-recomputed hash, and the hash inside the on-chain memo. If all three match, the decision is authentic and timestamped.
The on-chain memo format is AVOID.NET|v1|h:<b58-sha256>|d:<id>|t:<iso>

Find a signature on any investigation page's decision log, or run python -m src.verify_decision --signature <sig> for a CLI check.

Sequence
#2
Score
1818 (0)
Cluster
mainnet-beta
Slot
443522948
Off-chain at
2026-08-27T03:05:58.028Z
Anchored at
Block time

Independent verification

1. Database (off-chain)
HLnbJLB6i5zDDFRtHbcCVAD4VcgF1YbwofT6qn4xTtzA
2. Recomputed (your browser)
computing…
3. On-chain (Solana memo)
fetching…
Canonical bytes hashed (1389 chars)
{"actor":"reviewer","decided_at":"2026-08-27T03:05:57.904Z","decision":"review","investigation_id":"cbe21877-b631-47a7-a344-fe90e23b96f0","new_score":18,"page_slug":"bankr-bankrbot-ai-agent-prompt-injection-exploit","prev_score":18,"reason":"The page accurately establishes the core, well-corroborated facts of the May 4, 2026 Bankr/Grok incident -- a genuine prompt injection (attacker-controlled Morse-code text in Grok's public reply was treated by BankrBot as an authenticated command) compounded by an NFT-based privilege-escalation flaw that is properly a distinct 'over-permissioned agent' failure mode, and it appropriately surfaces remediation (Grok-reply disablement, IP whitelisting, reimbursement pledges). Its weakest point is asserting that the second, May 19-20 breach of 14 wallets used 'the same underlying permission-chain vulnerability' as the first, when contemporaneous sources describe that breach's mechanism as unconfirmed and possibly a private-key/session compromise rather than a repeat prompt-injection/NFT chain. A quote attributed to researcher 'Vadim' is accurate but not found in any of the three sources cited alongside it, and the DRB price-decline figure (15-20%) understates the ~40% decline reported elsewhere.","score_delta":0,"sequence_num":2,"submission_content_hash":null,"submission_id":null,"submission_kind":null,"submission_valence":null,"v":1}