paragraph-api

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs calling public Paragraph endpoints (e.g., https://public.api.paragraph.com/api/v1/posts, feed, and search) to fetch post and feed content from paragraph.com (user-generated/public content) which the agent is expected to read/process (includeContent/export flows) and could materially influence actions like publishing or sending emails.

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

• Direct money access detected (high risk: 1.00). The skill explicitly exposes crypto trading/execution endpoints. The SDK and REST docs include api.coins.buy and api.coins.sell (and /coins/buy, /coins/sell REST endpoints) that require a WalletClient and Account and return transaction hashes or build on-chain transactions. Those are explicit crypto/blockchain execution capabilities (signing/sending transactions), which constitute direct financial execution.