web3-viem

The provided skill file SKILL.md is purely informational and descriptive. It serves as documentation for the 'viem' TypeScript library, outlining its features, core concepts, utilities, and client interactions through code examples.

1. Prompt Injection: No patterns indicative of prompt injection (e.g., 'IMPORTANT: Ignore', 'Override', 'jailbreak') were found in the skill's content or metadata.
2. Data Exfiltration: The skill does not contain any commands or instructions that would lead to data exfiltration. While the code examples show network interactions (e.g., http() transport for blockchain clients), these are illustrative of the library's functionality and are not executed by the skill itself. Sensitive information like private keys and mnemonics are shown in examples (privateKeyToAccount("0x..."), mnemonicToAccount("legal winner thank year wave sausage ...")), but these are within code snippets for demonstration purposes, not actual commands for the agent to execute or hardcoded values used by the skill.
3. Obfuscation: No obfuscation techniques (Base64, zero-width characters, homoglyphs, URL/hex/HTML encoding) were detected.
4. Unverifiable Dependencies: The skill does not instruct the agent to install any external packages (e.g., npm install, pip install). It references mcp__context7__resolve-library-id and mcp__context7__get-library-docs, which are internal mechanisms for fetching documentation, not external code execution or dependency installation.
5. Privilege Escalation: No commands related to privilege escalation (e.g., sudo, chmod, service installations) were found.
6. Persistence Mechanisms: No commands attempting to establish persistence (e.g., modifying .bashrc, crontab) were found.
7. Metadata Poisoning: The skill's name and description fields are benign and accurately reflect its purpose.
8. Indirect Prompt Injection: As a documentation skill, it does not process external user input in a way that would make it susceptible to indirect prompt injection.
9. Time-Delayed / Conditional Attacks: No conditional logic for time-delayed or environment-specific attacks was found.

In conclusion, the skill is a static documentation resource. It does not contain any executable code or instructions for the AI agent, making it inherently safe from the identified threat categories.