Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add fancyboi999/goofish-cli --skill goofish-reply-buyergit clone --depth 1 https://github.com/fancyboi999/goofish-cliWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/fancyboi999/goofish-cli/goofish-reply-buyer)<a href="https://agentmods.dev/skills/fancyboi999/goofish-cli/goofish-reply-buyer"><img src="https://agentmods.dev/badge/skills/fancyboi999/goofish-cli/goofish-reply-buyer/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/fancyboi999/goofish-cli/goofish-reply-buyer"><img src="https://agentmods.dev/badge/skills/fancyboi999/goofish-cli/goofish-reply-buyer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00160 | $0.01727 |
| Opus 5 | $0.00080 | $0.00864 |
| Sonnet 5 | $0.00032 | $0.00345 |
| Haiku 4.5 | $0.00016 | $0.00173 |
Grade A, and why
goofish-reply-buyer scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 175 lines — stays where its author put it; the contents beside it link to each section on GitHub.
闲鱼回买家消息
工具命名
正文使用 auth_status 这类逻辑名。OpenClaw 调 goofish__<逻辑名>；
Claude Code / Cursor 调 mcp__goofish__<逻辑名>。frontmatter 的 allowed-tools
保留 Claude 权限预批准格式，OpenClaw 不读取该字段。
触发姿势
用户典型输入：
- "看看我未读消息"
- "有人问 XXX 怎么回"
- "那个议价的我回他多少"
- "催付一下下单没付钱的"
- "帮我批量处理一下消息"
核心原则（反复强调）
永远不自动发消息。
即使用户说"随便回一下"、"你看着办"，也要：
- 起草一条
- 展示给用户
- 等明确的"发 / 嗯发 / ok"后才调
message_send
为什么硬要这么做：
- 发错话在闲鱼成本高（外联词封号、情绪化引战、误报价）
- 买家一旦觉得"卖家不是本人"（AI 味太重），信任归零
- 和
message_watch的长连接事件流不同，单条对话是对外行为，不可逆
标准流程
Step 1 · 登录态自检
auth_status → valid=false 直接停，提示用户 auth login
Step 2 · 拉未读会话
message_list_chats(limit=20)
返回会话列表，每项含 cid / 对方昵称 / 最后一条消息预览 / 未读数。
筛选策略：
- 优先未读数 > 0 的
- 按最近活跃倒序
- 一次别处理太多（建议 ≤ 5 个会话一轮），消息太长用户看不过来
Step 3 · 逐会话拉上下文
对每个选中的 cid:
message_history(cid=..., limit=20)
20 条基本够分类意图了。太长的会话（50+ 条）拉 30 条。
同时（如果消息提到具体商品 ID）：
item_get(item_id=...) # 拿商品的最新状态，价格/标题/是否在售
Step 4 · 意图分类
详见 references/intent-classification.md。粗分 5 类：
| 意图 | 典型 | 处理策略 |
|---|---|---|
| 询价 / 基本信息 | "多少钱"、"还在吗"、"在哪发货" | 直接回，参考卖家商品描述 |
| 议价 | "便宜点"、"XXX 行吗" | 按 bargain-ladder.md 阶梯回 |
| 催付 | "已付款"、"明天到吗" | 回发货时间 |
| 售后 | "坏了"、"不对"、"要退款" | 谨慎回，留证据；严重时建议人工介入 |
| 外联 / 风险 | "加个微信"、"走转账" | block + 合规话术替换 |
Step 5 · 起草回复
按意图走不同子流程（各有对应 reference）。
起草完必过 risk-guard 扫描：
- 调用
goofish-risk-guard的external-contact-keywords.md扫草稿 - 命中 → 换词、重写或标红警告
Step 6 · 用户确认
展示格式：
【会话】买家昵称 "xxxxx"（未读 N 条）
【最新来信】
> "原文..."
【意图判定】议价（想砍到 XX 元，距你底线 XX 元）
【草稿回复】
"..."
【风控扫描】通过 ✅
发送这条吗？（y/改/跳过）
用户回 "发 / 嗯 / 是 / ok / 1" 才调 message_send。
用户回 "改" → 让他说具体改啥，重新起草。 用户回 "跳过" → 跳到下个会话。
Step 7 · 发送
message_send(cid=..., text="...")
成功后：
- 给用户确认"已发"
- 等一下再处理下一个（别连发多个 cid，触发频控）
卖家人设保持
Agent 要模仿用户自己的语气，不要用 AI 常见的书面语：
- ❌ "您好，非常感谢您的关注"
- ✅ "在的～"
- ❌ "经过慎重考虑，我们愿意给您一个优惠"
- ✅ "可以少 20，1200 行吗"
具体调教见 references/tone-guide.md。
特殊场景
场景：批量催付
用户说"把所有没付款的催一下"
list_chats找"已下单未付款"标签的会话- 每人起草一条温和催付话术
- 逐条确认（不要批量群发 — 给发送权等于关掉最后一道保险）
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 175 lines · 160 tokens per session scan A e5e5187dfbaf
goofish-reply-buyer is a skill published in the GitHub repository fancyboi999/goofish-cli (195 stars, last pushed 5d ago), licensed Apache-2.0. It adds 160 tokens to every session and 1,727 once invoked, about $0.0008 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
stripe
Generated by Gutenberg. Use this skill to call the stripe API through the local stripe CLI.
amazon-automation
Automate Amazon tasks via Rube MCP (Composio). Always search tools first for current schemas.
amazon-automation
Automate Amazon tasks via Rube MCP (Composio). Always search tools first for current schemas.
clishop
Search products, compare stores, place orders, and manage shopping flows through CLISHOP from OpenClaw.
baselinker-automation
Automate Baselinker tasks via Rube MCP (Composio). Always search tools first for current schemas.
bestbuy-automation
Automate Bestbuy tasks via Rube MCP (Composio). Always search tools first for current schemas.