FinProof
Adversarial AI Benchmark for Banking & Financial Services
Get picks like this daily. Today's launch numbers, the top launches, AI and tech news, and a daily opportunity pick, straight to your inbox.
Read the full description
Agent Siege: 6 agentic attacks against a banking AI — 60 seconds, no account. 11,474 adversarial prompts across 13 BFSI categories: • Indirect prompt injection via support ticket • MCP tool poisoning (metadata hijack) • RAG/memory poisoning • KYC bypass, PII exfiltration • Multi-hop traversal Mapped to RBI, SEBI, EU AI Act, SR 11-7, DPDP Act 2023. T1 public on HuggingFace. T4 withheld — no training on the test set.
Visit website ↗ Source listing ↗
PH · launched Oct 28, 2026 · 1 upvote (top 67%)
- Fintech
- Observability & eval
- B2B
- AI-native
- Commercial product
- Fintech
- Developer Tools
- Artificial Intelligence
Similar launches
See all 167 similar launches →Could you build this?
Partial A chat-based benchmark UI and test runner are easy to vibe-code. A credible banking-specific attack library and scoring methodology need deep BFSI and adversarial-ML expertise and curated domain data.
What it would actually take: Build a harness that sends prompts to target LLM or agent endpoints and scores responses with an automated judge. The hard part is authoring thousands of realistic, banking-specific adversarial cases mapped to regulatory frameworks, which requires security researchers and financial-compliance experts.
Competitors
Other products that read as similar to this one : 167 launches clear the similarity bar, closest 8 shown.
Attention rank: #90 of 164 (itself plus its competitors, highest first, normalized so YC and Product Hunt are compared fairly).
Launched 357 days after the earliest competitor.
- ACE · hn · 2026-04-05 · 9 upvotes · similarity 0.46
- jailbreak-archives · github · 2026-09-30 · 17 upvotes · similarity 0.43
- AgentGuard · ph · 2026-10-28 · 1 upvotes · similarity 0.42
- AgentExploitDB · ph · 2026-10-07 · 1 upvotes · similarity 0.41
- Agentwall · ph · 2026-09-09 · 2 upvotes · similarity 0.40
- We post-trained a model that pen tests instead of refusing · hn · 2026-06-20 · 93 upvotes · similarity 0.40
- ClawSecure · ph · 2026-03-15 · 316 upvotes · similarity 0.40
- ClawSecure · ph · 2026-05-11 · 247 upvotes · similarity 0.40