Hk pre owned rolex monitoring
Skill KelvinWu/hk-pre-owned-rolex-monitoring/skills/hk-pre-owned-rolex-monitoring
Platform-neutral Agent Skill for auditable Hong Kong pre-owned Rolex inventory monitoring.
npx -y skills add KelvinWu/hk-pre-owned-rolex-monitoring --skill hk-pre-owned-rolex-monitoringAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- 24 days oldThe repository was created 24 days ago. New is not bad, but a brand new repository carrying a familiar-sounding name is the shape a typosquat arrives in, and there has been no time for anyone else to find a problem with it.
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Monitor and explain Oriental Watch Hong Kong Rolex Certified Pre-Owned inventory with auditable history, retained listing images, and documented HK, APAC, Mainland China, or global market evidence. Use when a user asks whether 东方表行 or Oriental Watch Hong Kong pre-owned Rolex listings have new, removed, price, or detail changes; asks to review yesterday or a prior run; wants a delisted watch image; wants to compare an exact Rolex reference and nearby production years; or needs to initialize, diagnose, or repair this specific monitor. Do not use for generic inventory websites, new-retail Rolex pricing, or investment advice.
SKILL.md
5.2 KB, ~1.4k tokens by cl100k_base, as published. Nobody here has run it
HK Pre-owned Rolex Monitoring
监控香港东方表行 Rolex CPO，以可审计历史回答上新、下架、改价和资料变化，并在证据充足时解释香港、亚太、大陆或全球二手行情。库存事实与市场参考始终分开。所有确定性操作均调用 scripts/inventoryctl.py；不得在宿主提示词中重写基线、Diff、INVALID、行情聚合或通知幂等。
按意图路由
| 用户意图 | 操作 |
|---|---|
| 首次安装或 PyPI / PATH 故障 | 读取 references/host-compatibility.md，运行 scripts/bootstrap.py doctor；只有 INSTALL_VERIFIED 才算完成 |
| 创建监控 | 读取 references/skill-contract.md、adapter-contract.md、state-model.md；先运行 monitor init，确认问题后再 monitor create 和 monitor baseline |
| 查看今天或执行监控 | 运行 monitor status，再运行 monitor run；使用返回的 human_summary_zh、product_identity 和 next_actions |
| 查看昨天或历史结果 | 运行 monitor history 和 monitor show-run；不要重新抓取网站代替历史记录 |
| 生成用户报告 | 运行 report build --run-id <run-id>；需要行情时增加 --market-packet <file> |
| 行情对比 | 读取 references/source-access-policy.md 和 market-intelligence.md；先做来源诊断，再收集或构建 Market Packet |
| 通知 | 运行 outbox list；发送后用真实 provider、外部消息 ID 和核验结果执行 outbox ack |
| 修复、升级或恢复 | 先运行 monitor doctor；高风险操作前备份，只恢复校验通过的备份 |
CLI 每次只输出一个 JSON。优先执行回执中的 next_actions，但仍根据用户意图、宿主能力和是否需要用户确认选择下一步。
核心不变量
- 东方表行 stable ID 只能是
Lot_Number_Code；价格、图片、排序、时间和在售状态不得参与身份计算。 - 用户看到的每条变化必须包含商品名称、Rolex 型号和东方表行货号，不能只显示 stable ID。
- 首次基线不产生变化；INVALID、空结果、重复 ID 或部分抓取不得替换上一份成功快照。
- 图片存入
state-dir/images/<monitor-id>/，不存入对话或 Skill 目录；下架后保留历史图片。图片失败只产生警告。 - 行情失败、过期或证据不足不得修改库存基线；只有
benchmark_status=VERIFIED才能称为已验证参考价。 - 宿主任务和通知未经重新查询或真实回执验证，不得声称已经完成。
常见踩坑（Gotchas）
- 包索引首页能打开不代表依赖能够下载；必须让 Bootstrap 实际下载依赖并分别报告 DNS、TLS、
PIP_INDEX_URL/PIP_EXTRA_INDEX_URL、权限和 PATH。 stable_id是内部审计键，不是用户能感知的腕表名称；直接使用 CLI 返回的product_identity。- 商品链接和原图会下线；成功快照时就缓存图片，不要等到下架后再下载。
- 空列表或临时漏项不是“全部下架”；返回 INVALID 并保留成功基线。
- WatchCharts 型号估值没有生产年份，不能放入年份窗口 cohort；只能作为型号背景。
- 用户问“昨天”时读取保存的 run；重新运行得到的是今天，不是昨天。
outbox ack没有外部消息 ID或未经宿主复查时只能记为未验证，不能视为送达。
稳定入口
python scripts/inventoryctl.py skill info --json
inventoryctl skill info --json
inventoryctl skill self-test --json
inventoryctl monitor init --output <manifest> --json
inventoryctl monitor list --json
inventoryctl monitor history --id <id> [--date YYYY-MM-DD] --json
inventoryctl monitor show-run --run-id <run-id> --json
inventoryctl monitor run --id <id> --trigger <name> --json
inventoryctl market packet init|add|import-csv|attach-evidence|finalize ... --json
inventoryctl market compare --id <id>|--run-id <run-id>|--event-id <event-id> --file <packet> --json
inventoryctl report build --run-id <run-id> [--market-packet <packet>] --json
完整参数以 inventoryctl --help 为准。
按需参考
- CLI、结果、历史查询、报告和退出码：
references/skill-contract.md - Adapter 数据完整性：
references/adapter-contract.md - Runtime Plan 与宿主回传：
references/runtime-plan-contract.md - SQLite、图片、Outbox、备份和恢复：
references/state-model.md - Market Packet、年份匹配、聚合与报告：
references/market-intelligence.md - 来源权限、license 和时效门禁：
references/source-access-policy.md - 安装、镜像和 IDE / Agent 边界：
references/host-compatibility.md
Gives 0 of the 12 instructions most monitoring observability skills give in ~1.4k tokens
Counted across 481 of the 483 authors here whose files we hold, read 2026-08-06
- link every alert to a runbookin 43 of 481, across 35 files
- use structured json loggingin 36 of 481, across 31 files
- alert on user-facing symptomsin 20 of 481, across 15 files
- emit structured JSON logs with stable event namesin 18 of 481, across 13 files
- propagate trace context across boundariesin 16 of 481
- use histograms for latency trackingin 14 of 481, across 9 files
- use OpenTelemetry for distributed tracingin 13 of 481, across 8 files
- include a correlation ID on every log linein 13 of 481, across 8 files
- Define service level objectivesin 10 of 481, across 7 files
- Call useAzureMonitor before importing other modulesin 9 of 481, across 2 files
- stop and ask for clarification if inputs are missingin 9 of 481, across 2 files
- define on-call questions before adding telemetryin 9 of 481, across 4 files
Said here and by no other author read
- keep inventory facts separate from market references
- call the designated script for deterministic operations
- do not rewrite logic in the host prompt
- run the diagnostic tool for first setup or path failures
- initialize the monitor before creating a baseline
- run the monitor status command before a monitoring run
Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.