菜单
Deep shopping research skill for finding products similar to a reference link or image, benchmarking global pricing, comparing quality/provenance, and creating a reusable buying dossier with links, per-item prices, shipping notes, and value-focused recommendations.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
Brief pre-work alignment for any stage. Use before planning, coding, editing, reviewing, or delegating when the user asks you to understand the thesis, vibe, intent, current position, or direction of a request or existing thing. Returns 2–4 bold-labelled one-sentence observations and one short synthesis, then stops.
The single front door to the SDD pipeline (docs/plans/): research → plan → workshop → ADR → phase tasks → implement → progress → review → ship. Use when the user wants to plan, research, explore, specify or architect (now one atomic 'plan' step → one document: business spec + implementation plan), clarify, workshop, write an ADR, break a phase into tasks, implement or build a phase (optionally with a live review companion), update progress, code-review, or ship a plan (push + PR + watch checks) — or types /the-flow (or a legacy /plan-N command), or asks to start, resume, or adopt a plan flow. Guided mode (no args) coaches from durable on-disk state; direct jump runs one stage: /the-flow <id> <verb> [flags] — 1a explore, 1b plan (spec + plan in one doc; incl. clarify), 2c workshop, 3a adr, 5 tasks, 6 implement (--companion for live review), 6a progress, 7 review, 8 ship, 8c reconcile. Ids and verbs resolve alone; printed commands carry both.
Adaptive, high-signal post-action validation for whatever was just produced — unified plan, tasks, code change, workshop, legacy spec, or any artifact. Proves the claims that matter, reports only findings likely to change action, and adds independent critics only when risk or uncertainty earns them. Thesis- and consumer-aware; read-only through adjudication; compact verdict.
Pre-task token discipline. Use before any task when every read, tool call, line of reasoning, and output token must materially advance the stated goal or thesis, without weakening correctness or safety.
FlowSpace-first codebase research agent for parallel subagent exploration with structured output for parent orchestrator synthesis. Domain-aware v2.
Establish or refresh the project constitution and align the supporting norms documents before any planning phases begin. Re-entrant command that preserves user customizations during updates. Domain-aware v2.
| name | shopping-hunter |
| description | Deep shopping research skill for finding products similar to a reference link or image, benchmarking global pricing, comparing quality/provenance, and creating a reusable buying dossier with links, per-item prices, shipping notes, and value-focused recommendations. |
Use this skill when the user wants to find, benchmark, compare, or source products similar to a reference item, image, aesthetic, material, maker, or buying goal. The default mode is deep global shopping research with Australia prioritized, AUD-normalized pricing, and a documented dossier.
Before starting, ask for the missing items from this list. If the user already provided some, only ask for the missing/high-impact ones.
If the user says “don’t hold back,” “deep search,” or similar, proceed with a deep multi-pass workflow and do not over-ask beyond truly missing constraints.
Extract a target profile:
Separate:
When available, use parallel subagents or independent searches by scope. Good default passes:
Do not duplicate work already delegated to an agent. Wait for required research before final synthesis when feasible.
For each candidate, capture:
Never invent prices. If a price is from a snippet, archive, search result, or blocked page, label it clearly as unverified or partially verified.
If the user wants a file, create a Markdown dossier in the session files folder unless they explicitly request a repository path. If the user explicitly asks for a repo file, create it where requested.
Recommended dossier structure:
Use dense tables for product comparisons. Keep a short chat summary after creating the dossier and include the file path.
Be direct and opinionated:
Lead with the outcome:
<path>”Keep the chat response concise; the dossier carries the detail.