from scammer.skill62
Detect likely scam messages, identify the scam stage, and predict the scammer's next move to help users defend themselves.
This skill distills common scam playbooks into structured detection logic: it analyses user-provided messages or screenshots, identifies which stage of a scam flow the interaction is in, extracts key signals, and recommends next steps to stay safe. It is explicitly defensive — meant to help users recognise scams, not facilitate them.
Use this skill when a user pastes a suspicious message, uploads a screenshot of a conversation, or asks whether a contact or offer is legitimate. Triggers include commands like /detect-scam, /scammer [message], or natural-language prompts such as "帮我看看这条消息" / "这是诈骗吗". It also supports add/manage flows for maintaining a local scam pattern library.
This skill is written as a Claude/agent-style SKILL.md with instructions and tool calls (Read, Write, Edit, Bash). It will work best with agents that support file Read/Write and Bash tool integrations (Claude Code, agentic CLI-based agents).
This skill has not been reviewed by our automated audit pipeline yet.
