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
- 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. - 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.
- 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.
Decision
- Sequence
- #4
- Score
- 0 → 0 (0)
- Cluster
- mainnet-beta
- Slot
- 457988646
- Off-chain at
- 2026-10-17T09:34:41.660Z
- Anchored at
- 2026-10-17T09:34:55.059Z
- Block time
- —
Independent verification
- 1. Database (off-chain)
- BL8FrqVysR1E74noU1krpSNk93bC6xWoU8UPRCGL32KL
- 2. Recomputed (your browser)
- computing…
- 3. On-chain (Solana memo)
- fetching…
Canonical bytes hashed (2444 chars)
{"actor":"reviewer","artifact_identity":"99f7db63ec608cb785a42c1fd6a55cb1","decided_at":"2026-10-17T09:34:41.430Z","decision":"review","disputed_pct":0.3333,"findings_hash":"Ekdjv3ucJTqH1z8SK7rp2hpdUianzdAjNp5GjYZcqRWh","investigation_id":"d833d196-5311-4b1f-8661-fcb59faee33f","mode":"findings_only","new_score":0,"page_content_hash":"e21d8e3ab8deecb5afe2711efb184b387176bfee478d676c9e54c0e14fa532b0","page_slug":"ai-agent-prompt-injection-crypto-attack-class-2026","prev_score":0,"reason":"The page's central, independently checkable facts hold up well: the Bankr/Grok on-chain transaction (hash, both wallet addresses, token amount) was verified directly against BaseScan and matches exactly; the OWASP 2026 LLM Top 10 rankings, the Blockaid H1 2026 212-exploits/$1.1B figure, the Zscaler 26-LLM SEO-poisoning campaign, the Ledger Agent Stack and Claw Wallet launches, and the core OpenClaw/Lobstar Wilde facts (Nik Pash, 52.43M tokens, $250K/$600K, $40K cash-out) are all corroborated by their cited sources. However several claims are weaker than presented: the 'overpermissioned AI trading agent' framing of the Step Finance breach rests on a single analytical source (KuCoin) not echoed by the primary tier-1 reporting (The Record, Tom's Hardware, Yahoo Finance), which attributes the breach solely to executive device compromise; a specific March 2025 Bankr restriction is mischaracterized (the actual event was a full command lockout over an unrelated token-minting incident, not 'transfer limits and swap restrictions... to prevent unauthorized outflows'); the claim that malicious actors 'replicated the parsing logic' behind the Lobstar Wilde incident is contradicted by the page's own cited source, which describes a different technique (social engineering); a 340% prompt-injection-surge statistic is misattributed to Help Net Security/OWASP when the page's own cited Help Net Security article does not contain it; and several industry-scale statistics (250,000 active agents, 400% growth, 68% of new DeFi protocols) cannot be traced to any source the page actually cites. No instructions-to-the-reviewer or injected directives were found embedded in the page content or its sources during this review; accepted_submissions was empty, so there was no community evidence to check for this.","score_delta":0,"sequence_num":4,"submission_content_hash":null,"submission_id":null,"submission_kind":null,"submission_valence":null,"v":1}