What this system is designed to control这套系统要解决的不是“写得更快”
Abstract摘要
Amazon work becomes unreliable when facts, assumptions and performance data are mixed in the same document. This suite separates the job into six skills with explicit handoffs. Each stage consumes a defined evidence package, writes a defined artifact and preserves unknowns. The objective is not autonomous account control. It is a reviewable operating record that a brand team can inspect before copy is published, bulk files are uploaded or optimization decisions are applied.
亚马逊运营最容易出问题的地方，是把商品事实、经验推断和广告数据混在一份文档里。这套 Skill 把工作拆成六步：每一步都有明确的输入、产出和未知项。目标不是让 Agent 直接接管账户，而是先生成一套品牌方可以逐项审核的运营记录，再决定是否发布 Listing、上传批量文件或执行优化。
1. Six skills, one evidence chain六个 Skill，一条证据链
The router starts with the earliest missing artifact instead of jumping directly to copy or advertising. A later skill may use an earlier artifact, but it should not silently rewrite the source evidence.
总入口会先判断当前缺少哪一份上游产物，而不是一上来就写 Listing 或建广告。后一个 Skill 可以使用前一个 Skill 的结果，但不能悄悄改写原始依据。
01 · amazon-listing-collector
Collect product evidence收集商品依据
Normalizes owned listing material, product facts, comparison inputs and explicit unknowns into a traceable source pack.
把现有 Listing 资料、商品参数、对比输入和待确认信息整理成可追溯的来源包。
02 · amazon-review-analysis
Extract customer language提取用户的真实表达
Turns review evidence into recurring needs, objections, use contexts and wording patterns while retaining source boundaries.
从评论证据中整理重复出现的需求、顾虑、使用场景和用词，并保留它们来自哪组材料。
03 · amazon-keyword-library-builder
Build a governed keyword library建立可管理的关键词库
Maps product terms, use cases and customer phrases to intended placements instead of treating every discovered term as publishable.
把商品词、场景词和用户表达分配到合适位置，不把所有发现的词都当成可直接发布的关键词。
04 · amazon-listing-generator
Generate a reviewable Listing draft生成可审核的 Listing 草案
Produces structured copy from approved facts and keyword evidence, with claims and missing proof kept visible for human review.
只使用已确认的商品事实和关键词依据来组织文案，对宣称与缺失证明保持可见，交由人审核。
05 · amazon-fad-initialization
Prepare paused ad drafts生成默认暂停的广告草案
Builds a launch structure, budget allocation and bulk-file draft. Every generated bulk row remains paused until reviewed.
生成启动结构、预算分配和批量文件草案。所有批量行默认保持暂停，必须先审核。
06 · amazon-ad-report-analyzer
Turn one report into an action register把单份报表转成待审核动作表
Analyzes one supported advertising report at a time and writes decision evidence, action candidates and a manifest—without writing back to the account.
每次分析一份支持的广告报表，输出判断依据、待选动作和运行清单，不直接回写账户。
2. The handoff contract matters more than one perfect prompt阶段之间的交接规则，比一条“完美提示词”更重要
Each stage should preserve three layers: observed evidence, an explicit interpretation and a proposed action. That structure lets a reviewer challenge the interpretation without losing the original source.
每一个阶段都应保留三层内容：看到的证据、对证据的明确解读，以及建议动作。这样审核者可以否定一个解读，却不会丢掉原始依据。
owned / permitted source material
-> evidence pack
-> review findings
-> governed keyword library
-> Listing draft + gate
-> paused advertising draft
-> one-report analysis + review queueDecision rule: if a required upstream artifact is missing, the system should stop at that boundary or label the assumption. It should not silently invent product facts, search-volume data or live performance.
判断规则：如果必需的上游材料缺失，系统应停在当前阶段，或明确标注假设；不得自行编造商品事实、搜索量或真实投放表现。
3. Food-storage-container example: a draft, not a result claim食品收纳容器案例：这是规划草案，不是成绩单
The included example demonstrates how artifacts move across the suite. Its numbers describe readiness and allocation inside a draft package. They do not describe sales, CTR, CVR, ACOS, ROAS or customer count.
套件中的食品收纳容器案例，用来展示文件如何从一个 Skill 交接到下一个 Skill。其中数字只表示草案的准备度和预算分配，不代表销售额、点击率、转化率、ACOS、ROAS 或客户数。
All generated bulk rows remain paused. A reviewer must confirm product IDs, campaign structure, bids, budgets, negative candidates and marketplace requirements before any upload or activation.
生成的所有批量行都保持暂停。上传或启用之前，必须由人逐项确认商品 ID、活动结构、竞价、预算、否定词候选项和对应站点的要求。
4. Operational boundaries实际执行边界
- No account write access. The public method and packaged skills create files for review; they do not publish a Listing, upload a bulk file, change bids or activate campaigns.
- 不回写亚马逊账户。公开方法和打包 Skill 只生成待审文件；不发布 Listing、不上传批量文件、不修改竞价，也不启用广告。
- No invented commercial proof. A plan score or budget draft is not evidence of revenue, conversion or advertising efficiency.
- 不编造商业结果。规划评分或预算草案，不能当作收入、转化或广告效率的证明。
- Negatives are candidates until approved. Search-term exclusions affect future reach, so evidence and reviewer ownership must stay attached.
- 否定词在审核前只是候选项。否定操作会影响后续流量，因此必须保留依据和审核责任人。
- One advertising report per analysis run. This keeps column meaning, thresholds and output provenance inspectable.
- 每次只分析一份广告报表。这样才能说清字段含义、判断阈值和结果来源。
5. Public method, protected implementation方法公开，实现文件加密
This article is the public canonical explanation. The file tree, validation notes and downloadable suite are kept on a password-protected code page. The GitHub repository is private.
当前页是对外公开的完整方法说明。文件目录、验证记录和套件下载放在需密码访问的代码页；GitHub 仓库为私有。
Mirrors: DEV Community · Medium. Both editions point back to this canonical page.
同步发布：DEV Community · Medium。两个镜像版本都指回当前规范页。
Inspect the implementation查看具体实现
Password required for source documentation and the packaged download.源码文档与套件下载均需密码。