Wallets

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

• Third-party content exposure detected (high risk: 0.90). The required workflow runs curl -s https://ethskills.com/wallets/SKILL.md and the agent reads that fetched page text into its LLM context; this is outsider-authored free text from a public web source at runtime.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The catalog explicitly instructs runtime fetching and agent consumption of remote instructions via "curl -s https://ethskills.com/wallets/SKILL.md" and "Read https://ethskills.com/wallets/SKILL.md before building", so the external URL https://ethskills.com/wallets/SKILL.md is a required runtime dependency that directly controls agent prompts.

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

• Direct money access detected (high risk: 1.00). The skill is explicitly named "Wallets" and provided by "EthSkills" (Ethereum-themed). That naming indicates a crypto wallet capability (managing wallets, signing/sending transactions, swaps), which is a specific financial execution tool rather than a generic utility. Even though the file is short, the explicit wallet/crypto context and external install imply the agent would have direct ability to perform blockchain financial operations (sign/send transactions), so it grants Direct Financial Execution Authority.