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 xjli360/sealeap-amazon-ad-skills --skill sealeap-xiezhi-amazon-ad-relevance-weight-bootstrapgit clone --depth 1 https://github.com/xjli360/sealeap-amazon-ad-skillsWrote 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/xjli360/sealeap-amazon-ad-skills/sealeap-xiezhi-amazon-ad-relevance-weight-bootstrap)<a href="https://agentmods.dev/skills/xjli360/sealeap-amazon-ad-skills/sealeap-xiezhi-amazon-ad-relevance-weight-bootstrap"><img src="https://agentmods.dev/badge/skills/xjli360/sealeap-amazon-ad-skills/sealeap-xiezhi-amazon-ad-relevance-weight-bootstrap/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/xjli360/sealeap-amazon-ad-skills/sealeap-xiezhi-amazon-ad-relevance-weight-bootstrap"><img src="https://agentmods.dev/badge/skills/xjli360/sealeap-amazon-ad-skills/sealeap-xiezhi-amazon-ad-relevance-weight-bootstrap.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00062 | $0.01186 |
| Opus 5 | $0.00031 | $0.00593 |
| Sonnet 5 | $0.00012 | $0.00237 |
| Haiku 4.5 | $0.00006 | $0.00119 |
Grade A, and why
sealeap-xiezhi-amazon-ad-relevance-weight-bootstrap 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 5d 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.
What it actually says
Amazon 广告相关性启动
目标
用精准购物意图、页面一致性和受控学习预算积累可解释的点击与转化信号，而不是依赖未经证实的权重公式。
适用任务
- 为新品设计前几天的广告学习计划。
- 诊断相同出价但曝光差异。
- 在提高竞价前检查关键词和 Listing 相关性。
开始前要拿到
- 目标 ASIN/SKU、站点和产品事实。
- 精准词、搜索结果相关性和 Listing 字段覆盖。
- 建议竞价、广告位、预算、盈亏 CPC 与止损。
- CTR、CVR、CPC、订单和归因窗口基线。
缺少字段时列出证据缺口，并把相关结论标为 FACT、ESTIMATE、ASSUMPTION 或 UNKNOWN；不要补造数据。
不可妥协的边界
- 第三方数据均为估算或代理证据；Amazon 一方报告、后台实时字段和产品事实优先。
- 经验阈值只能作为可调起点，必须展示敏感性分析，不能写成 Amazon 官方规则。
- 不得捏造销量、搜索量、CPC、CVR、成本、认证、产品属性或消费者需求。
- 默认提供诊断或草案。写入前展示对象、旧值、新值、影响、停止线与回退，核对用户已有授权是否覆盖对象、动作与预算；范围已明确授权时继续执行并回读核验，只有未覆盖或扩大的范围才请求批准。
- 不输出或保存素材来源身份、账号、链接、作品编号、互动数据、原始话术或其他可反查来源的线索。
- 广告写操作须在用户明确授权的对象、动作与预算范围内，并保留回退值；已有授权覆盖时不重复索取。
- 不得把经验性拍卖解释或短期相关性写成 Amazon 官方公式。
工作流
1. 建立相关性地图
将产品核心属性、对象和场景与搜索词、Listing 标题/要点和目标页面逐项对齐。
2. 选择启动词
优先购买意图明确且搜索结果高度一致的词，排除大而泛、与产品弱相关的流量。
3. 设置学习护栏
以建议竞价区间为参考制定小范围测试，预先限定日预算、累计花费、最低样本和停止条件。
4. 观察分层信号
分别看曝光、CTR、CVR、CPC、广告位和搜索词；先判断资格/相关性，再判断商品页和价格。
5. 逐步调整
一次只改变主要变量；只有转化证据支持时扩大预算或竞价，表现恶化则回退。
判断标准
- 广告排序受出价、相关性、预计效果和竞争环境共同影响；不存在可直接读取的固定单一权重分。
- 新品可进行受控学习，但不得无上限高价抢位或把前三天当作必然起量窗口。
- 盈亏 CPC = 广告前每单贡献毛利 × 订单 CVR；订单 CVR 按归因订单 / 点击，以小数代入。出价与实际 CPC 分开，并检查动态竞价和位置调整后的风险。
第三方 MCP 数据
需要外部关键词、竞品、评论或公开网页证据时，读取 references/mcp-data-plan.md，并使用 scripts/mcp_research.py。
- 先动态执行
tools/list、search-tools和describe，依据实时inputSchema构造参数。 - 凭证只从环境变量读取，不进入参数、URL、Skill、终端输出或 Git。
- 可能计费的
tools/call先展示 Provider、工具、无密钥参数、预计成本与输出位置，核对已有授权；仅在授权覆盖本次范围时使用--allow-cost，该标志不是费用上限。 - 脱敏结果用
--output写入 Skill 包之外的任务私有目录；不假设安装位置受仓库.gitignore保护。第三方数据标为估算或代理证据。 - 失败一次后记录缺口，不以重复付费重试掩盖不可用状态。
必须交付的结果
- 关键词—页面相关性矩阵
- 启动广告草案
- 预算与累计止损
- 学习期监控表
- 待批准最小变更集
结尾列出站点、数据窗口、证据来源、关键假设、缺口、风险、下一步和所有待批准动作。证据不足时写 HOLD，不得包装成可直接执行。
执行细节、证据字段和质量检查见 references/playbook.md。
What ships with it
4 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.
- 5d ago First seen · 86 lines · 62 tokens per session scan A 110f21275d2c
sealeap-xiezhi-amazon-ad-relevance-weight-bootstrap is a skill published in the GitHub repository xjli360/sealeap-amazon-ad-skills (86 stars, last pushed 6d ago), licensed MIT. It adds 62 tokens to every session and 1,186 once invoked, about $0.0003 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 2026-09-07.
Other skills, from other repositories
zach-seller-skill-creator
A Chinese-language guide for Amazon sellers who want to turn repeated work processes into reusable skills for an AI agent.
zach-search-term-analyzer
An analyzer for Amazon Brand Analytics Top Search Terms reports, which show popular searches across Amazon and how clicks and conversions are distributed among products.
zach-search-term-report-analyzer
An Amazon Ads search-term report analyzer for Sponsored Products, Sponsored Brands, and Sponsored Display campaigns. It groups related search terms, measures results over 7, 14, and 30 days, and produces reports in several file formats.
amazon-analysis
Amazon-domain general analysis and multi-endpoint research engine. Handles broad or composite Amazon research requests that span multiple data dimensions or have no single specialized angle. Use when: - user asks for multi-endpoint Amazon research, composite reports, or general Amazon market/product analysis user asks…
amazon-market-trend-scanner
Amazon category trend scanner. Scans Amazon category landscapes to discover trending subcategories, emerging niches, and market shifts. Tracks demand surges, brand consolidation, new entrant waves, price band migration, and margin changes across all subcategories under a parent category. Use when user asks about…
amazon-pricing-command-center
Data-driven pricing strategy engine for Amazon sellers. Given one or more ASINs, auto-detects each product's leaf category, analyzes the pricing landscape, and delivers RAISE/HOLD/LOWER signals with profit simulation. Supports single ASIN or batch (multiple ASINs, auto-grouped by category). Uses ZooData API endpoints…