emblem-ai-prompt-examples

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.80). The skill's main documentation and reference files (e.g., references/market-research.md and references/nft-opensea-examples.md) include numerous required example prompts that instruct the agent to fetch and summarize untrusted public web sources (e.g., "Get the tokenomics details from Jupiter's website", "Pull the top performing pools from DeFiLlama's DEX page", "Get OpenSea stats for this collection slug", and "Summarize current social narrative ... from recent sources"), which clearly exposes the agent to arbitrary third‑party content that could influence subsequent decisions or tool use.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill is explicitly focused on financial/crypto workflows: it includes wallet operations, portfolio review, trading and DeFi (quote-first swaps, routing, yield), transfers, cross-chain and conditional orders (stop-loss, take-profit), Bitcoin ordinals, prediction-market (Polymarket) order prompts, NFT marketplace flows, and Emblem Vault minting/key-reveal. These are specific crypto/blockchain and market-order related capabilities (not generic tooling), so it is specifically designed for financial execution workflows.