The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the AI agent to use curl to interact with a local service at 127.0.0.1:18799. This includes health checks and manual HTTP requests to the /mcp endpoint to execute wallet operations.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests and processes data from the local wallet service (such as tool definitions and wallet query outputs) which could theoretically be manipulated to contain adversarial instructions. Ingestion points: Data returned from the tools/list and wallet-specific tool calls. Boundary markers: No specific delimiters or safety instructions are provided to distinguish between data and commands in the tool outputs. Capability inventory: The agent can execute shell commands via curl and perform network operations on localhost. Sanitization: There is no evidence of sanitization or validation of the content retrieved from the local MCP server before it is processed by the agent.

superise-mcp-usage