The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The documentation in SKILL.md and references/react-integration.md instructs users to download and install the @inferencesh/sdk package from the public NPM registry.
[COMMAND_EXECUTION]: Documentation in references/agent-patterns.md and references/tool-builder.md describes how to enable the SDK's built-in code execution capability for agents using the internalTools().codeExecution(true) method.
[DYNAMIC_EXECUTION]: Educational code examples in references/agent-patterns.md and references/tool-builder.md demonstrate the use of eval() to process mathematical expressions provided by an agent through a tool call.
[PROMPT_INJECTION]: The skill architecture handles untrusted user input via the agent.sendMessage function, as documented in SKILL.md and references/react-integration.md. Ingestion points include direct user messages and file attachments. Capability inventory for agents includes tool execution (e.g., the delete_file example in references/agent-patterns.md), server-side code execution, and network operations via webhook tools. The provided examples do not explicitly demonstrate boundary markers or sanitization for input data.

javascript-sdk