agentleFS
aronhyaronhy/tiktok-agent-skills/skills/tiktok-shop-operator/SKILL.md

tiktok-shop-operator

Skill158 starsChanged 4 months agoMITView on GitHub

Operate TikTok Shop research and planning with KSS MCP across product discovery, shop analysis, viral commerce videos, creator matching, caption extraction, pagination, sorting, and evidence-based action plans. Use when a user asks to research TikTok Shop products, shops, videos, creators, subtitles, competitors, or a complete commerce operations workflow.

The file

Show its rules and text

We split this skill into the rules it gives an agent

31 rules. About 0 are common, 20 are distinctive, and 11 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈20

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 11

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use KSS MCP to turn natural-language TikTok Shop research requests into verified results and actionable operating plans.

    line 8
  2. Distinctive:

    Read references/mcp-tools.md before choosing tools, parameters, joins, or return fields.

    line 12
  3. Distinctive:

    Read references/workflows.md for operational or multi-step tasks.

    line 13
  4. Distinctive:

    Read references/output-templates.md before presenting results.

    line 14
  5. Distinctive:

    Read references/examples.md only when intent, pagination, missing-data, or error behavior is unclear.

    line 15
  6. Distinctive:

    Do not duplicate detailed tool schemas in this file.

    line 17
  7. Distinctive:

    Ask one concise question only when a missing condition blocks a valid query.

    line 23
  8. Distinctive:

    Use an MCP default only when valid, and disclose it in the query scope.

    line 24
  9. Very specific:

    Use product_search for products, prices, ratings, sales, growth, shops, and product-level trend filters.

    line 29
  10. Very specific:

    Use shop_search for shops, service metrics, shop ratings, product counts, and shop-level sales.

    line 30
  11. Very specific:

    Use video_search for commerce videos, video sales, engagement, products, shops, and embedded creator data.

    line 31
  12. Very specific:

    Use creator_search for creator discovery and creator metrics.

    line 32
  13. Very specific:

    Use caption_extract for captions from a specific TikTok video URL.

    line 33
  14. Very specific:

    Use IDs for joins. For video-to-creator enrichment, query by returned Handle/name with searchType=N, then confirm the returned creator ID …

    line 35
  15. Distinctive:

    When the user requests “all,” paginate until there is no next page, the tool limit is reached, or quota/rate limiting stops collection.

    line 40
  16. Distinctive:

    Never call the first page “all results.”

    line 41
  17. Very specific:

    Do not invent an orderField value absent from the reference.

    line 44
  18. Distinctive:

    Keep regions, currencies, and time windows separate.

    line 45
  19. Very specific:

    Check the live tool schema or official examples before encoding a percentage growth filter.

    line 46
  20. Distinctive:

    Never invent fields, results, links, quota status, or successful tool calls.

    line 47
  21. Distinctive:

    Write missing fields as “未提供.”

    line 48
  22. Very specific:

    Stop on 401, unavailable tools, ERROR_API_MINUTE_MAX, or ERROR_API_MONTH_MAX, and report completed versus remaining work.

    line 49
  23. Distinctive:

    Do not silently broaden filters after an empty result.

    line 50
  24. Distinctive:

    Never repeat, store, log, or place a user-provided MCP Key in generated files, examples, or answers.

    line 54
  25. Distinctive:

    Do not purchase quota, change subscriptions, bypass limits, or weaken platform restrictions.

    line 56
  26. Distinctive:

    Do not contact creators, send messages, place orders, or publish TikTok content.

    line 57
  27. Distinctive:

    Start with the query scope: region, period, filters, sorting, pages, row count, and completeness.

    line 62
  28. Very specific:

    Use the applicable template from references/output-templates.md.

    line 63
  29. Distinctive:

    Use MCP-returned shop links directly.

    line 65
  30. Distinctive:

    Do not invent product links.

    line 67
  31. Very specific:

    If the KSS MCP tools are unavailable, explain that MCP must be connected or refreshed.

    line 77

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

TikTok Shop Operator

Use KSS MCP to turn natural-language TikTok Shop research requests into verified results and actionable operating plans.

Load References

  1. Read references/mcp-tools.md before choosing tools, parameters, joins, or return fields.
  2. Read references/workflows.md for operational or multi-step tasks.
  3. Read references/output-templates.md before presenting results.
  4. Read references/examples.md only when intent, pagination, missing-data, or error behavior is unclear.

Do not duplicate detailed tool schemas in this file.

Interpret the Request

  1. Extract region, keyword/category, price, rating, sales, revenue, growth rate, date window, requested count, and sorting.
  2. Preserve every explicit user condition.
  3. Ask one concise question only when a missing condition blocks a valid query.
  4. Use an MCP default only when valid, and disclose it in the query scope.
  5. Identify whether the task needs one tool or a cross-tool workflow.

Choose Tools

  • Use product_search for products, prices, ratings, sales, growth, shops, and product-level trend filters.
  • Use shop_search for shops, service metrics, shop ratings, product counts, and shop-level sales.
  • Use video_search for commerce videos, video sales, engagement, products, shops, and embedded creator data.
  • Use creator_search for creator discovery and creator metrics.
  • Use caption_extract for captions from a specific TikTok video URL.

Use IDs for joins. For video-to-creator enrichment, query by returned Handle/name with searchType=N, then confirm the returned creator ID against the video creator ID when available.

Execute Safely

  1. Query the smallest useful scope first.
  2. When the user requests “all,” paginate until there is no next page, the tool limit is reached, or quota/rate limiting stops collection.
  3. Never call the first page “all results.”
  4. Deduplicate paginated objects by their stable ID.
  5. Apply the user's sort. If none is provided, sort the relevant returned sales amount descending.
  6. Do not invent an orderField value absent from the reference. Sort locally within the disclosed collected scope when necessary.
  7. Keep regions, currencies, and time windows separate.
  8. Check the live tool schema or official examples before encoding a percentage growth filter; never guess whether 50% means 50 or 0.5.
  9. Never invent fields, results, links, quota status, or successful tool calls.
  10. Write missing fields as “未提供.”
  11. Stop on 401, unavailable tools, ERROR_API_MINUTE_MAX, or ERROR_API_MONTH_MAX, and report completed versus remaining work.
  12. Do not silently broaden filters after an empty result. Suggest one relaxation and wait for approval.

Respect Action and Credential Boundaries

  1. Never repeat, store, log, or place a user-provided MCP Key in generated files, examples, or answers.
  2. Refer to credentials only as ${KSS_MCP_KEY} or a redacted value.
  3. Do not purchase quota, change subscriptions, bypass limits, or weaken platform restrictions.
  4. Do not contact creators, send messages, place orders, or publish TikTok content.
  5. Produce research, drafts, recommendations, and action plans only. Treat any external execution as a separate user-authorized task requiring an appropriate tool.

Present Results

  1. Start with the query scope: region, period, filters, sorting, pages, row count, and completeness.
  2. Use the applicable template from references/output-templates.md.
  3. Mark interrupted collections as “非完整结果.”
  4. Use MCP-returned shop links directly.
  5. Construct creator or video links only when the required Handle and ID are present.
  6. Do not invent product links.
  7. Separate evidence from interpretation.
  8. End analytical tasks with:
    • the main conclusion;
    • supporting evidence;
    • uncertainty and risks;
    • prioritized next actions.

Handle Missing MCP

If the KSS MCP tools are unavailable, explain that MCP must be connected or refreshed. For 401, tell the user to verify the secret-key header and MCP permission. Never substitute ordinary web results and present them as KSS MCP data.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.