The Agent Skills Directory

[SAFE]: The skill is purely instructional and contains no executable scripts or commands. It functions as a safety manual for an agent interacting with the X Money service.
[SAFE]: Robust security guardrails are established, including a requirement for explicit user approval for every financial action (sending, requesting, or spending money) and strict prohibitions against asking for or displaying sensitive information such as passwords, passkeys, or credit card details.
[SAFE]: Network references are limited to well-known, official domains associated with the X (formerly Twitter) platform (e.g., x.com, money.x.com), and the skill correctly identifies these as the authoritative sources for connection management.
[INDIRECT_PROMPT_INJECTION]: The skill defines an attack surface by instructing the agent to relay 'message' content received from external MCP tool outputs. This finding is categorized as safe because the skill provides explicit instructions to the agent to interpret and relay these messages in its own words and to adhere to the strict financial guardrails defined in the guidance.
Ingestion points: SKILL.md specifies that the agent should relay the message field from X Money MCP tool results.
Boundary markers: No explicit delimiters are defined for the relayed strings.
Capability inventory: The skill involves high-privilege financial capabilities including sending money and merchant payments.
Sanitization: No technical sanitization is mentioned, though the skill provides clear behavioral constraints for the agent.

cursor-x-money-guide