TikTok Content Creation Compliance

Review TikTok Shop seller and creator promotional content compliance using official TikTok Shop Seller University guidance. Use when an AI agent needs to che...

Security Scan

The skill is an instruction-only TikTok Shop compliance reviewer whose requirements and instructions are consistent with its stated purpose and do not request unexpected credentials or installs.

Purpose & Capability

The name/description (TikTok Shop content compliance) matches the instructions and provided reference file. The skill asks the agent to review scripts, videos, captions, listing images, and to consult TikTok Shop guidance — all coherent with the declared purpose. No unrelated binaries, credentials, or config paths are requested.

Instruction Scope

The SKILL.md directs the agent to inspect user-provided media (video files, frames, metadata, subtitles) and to consult the bundled reference summary; it also tells the agent to reopen the official TikTok URLs when the user requests the latest rules. Inspecting provided media and fetching live policy pages is reasonable for this task, but it means the agent will access user media and make outbound network requests to the listed TikTok URLs when asked to verify live policy.

Install Mechanism

No install spec or code is included (instruction-only). Nothing will be downloaded or written to disk by the skill itself during install.

Credentials

The skill requires no environment variables, credentials, or config paths. The instructions do not request secrets or unrelated service tokens.

Persistence & Privilege

Skill flags are default (always: false, user-invocable: true). It does not request permanent presence, nor does it instruct modifying other skills or system-wide settings.

Assessment

This is an instruction-only reviewer that uses the bundled policy summary and (on demand) the official TikTok Seller University pages. Before installing or invoking it: 1) be prepared to supply any video or listing media you want reviewed — the agent will inspect those files; 2) understand the agent may fetch the official TikTok URLs to verify live policy when you ask for the latest rules (network access); 3) do not provide unrelated private credentials or secrets to this skill (it does not need them); and 4) if you rely on the bundled reference file, remember it reflects policy as of 2026-03-20 and you should ask the agent to fetch live pages for time-sensitive cases.

Overview

Classify TikTok Shop promotional content as Permitted, Risky - revise, Prohibited, or Unclear - verify live policy. This skill applies to both sellers promoting their own products and creators promoting products for others, including affiliate creator workflows.

Use the official policy summary in references/policy-summary.md as the starting point, and stay conservative when a claim, visual, product anchor, promotion mechanic, or edit could mislead users.

If the user asks for the latest or current rule, reopen the official URLs in the reference file before giving a final answer. TikTok Shop policy guidance changes over time.

Source hierarchy and dedupe

Use references/policy-summary.md as the canonical rule map, not as a page-by-page dump.

When multiple TikTok Shop pages repeat the same idea:

• prefer the most specific policy page as the primary source
• use broader "best practices" or "policy pro" pages as reinforcement, examples, or clarifications
• summarize repeated rules once instead of restating them under multiple headings
• if two pages appear to conflict, prefer the newer and more specific page, then mark the case Unclear - verify live policy if the conflict is still unresolved

Practical order of authority:

1. Product-category, health, AI, giveaway, gambling, platform-safety, and affiliate-eligibility pages
2. Core content-policy and enforcement pages
3. "Best Practices", "Creating with Impact", "Avoid Misleading Content", beauty and skincare guidance, and other "Become a Policy Pro" explainer pages

Review Workflow

1. Identify the artifact

Determine what you are reviewing:

• video concept
• script
• shot list
• caption
• product claims
• editing plan
• finished video description
• exported video file
• product listing image or PDP image

Also determine the promotion context:

• seller promoting own product
• creator promoting own shop product
• affiliate creator promoting another seller's product
• creator asking whether they are even eligible to promote the selected category

If the user only provides a rough idea, infer the missing structure and review the idea anyway.

If the user provides a video file, inspect the actual media instead of relying only on the script or filename. Prefer:

• video metadata such as duration, aspect ratio, and subtitle tracks
• sampled frames across the full runtime
• on-screen text or subtitle extraction when available
• comparison against the linked product image or listing media

2. Check the highest-risk areas first

Review in this order:

1. Product-category or affiliate eligibility limits
2. Health, wellness, medical, fertility, pregnancy, sexual wellness, weight-loss, or muscle-gain claims
3. AI-generated or AI-edited people, product visuals, disclosures, or expert personas
4. Misleading product presentation, unrealistic effects, before-and-after logic, or exaggerated promises
5. Still-frame, low-quality, pre-recorded, or non-interactive video or LIVE formats
6. Gambling, giveaway, or purchase-incentive mechanics
7. Off-platform traffic redirects, IP, counterfeit, knockoff, or unoriginal-content risks
8. Nudity, sexualized behavior, minors, violence, harassment, or shocking-content issues

Load references/policy-summary.md before deciding.

If the user is working from a creator education page, treat that page as a shortcut into the relevant canonical rule. Do not overcount duplicate warnings just because the same issue appears in several training pages.

When reviewing a finished video, explicitly check:

• whether the shown package, bottle count, size, colorway, or bundle format matches the listing image
• whether watermarks from external tools, stock platforms, or AI systems are visible
• whether the edit appears fully AI-generated or significantly AI-edited, triggering disclosure review
• whether subtitles or dialogue add claims that are weaker or stronger than the visuals

3. Produce a policy verdict

Use exactly one primary verdict:

• Permitted
• Risky - revise
• Prohibited
• Unclear - verify live policy

For every flagged issue, name:

• the risky claim, visual, or edit
• why it is risky
• the matching policy section from the reference file
• the official source URL

4. Rewrite toward compliance

When revising content:

• replace hard promises with neutral, support-style wording
• align every claim with the actual listing, packaging, and product capabilities
• remove any fake authority signals, miracle framing, or fear tactics
• suggest AI disclosure when content is fully AI-generated or significantly AI-edited
• remove off-platform redirects, random-chance mechanics, or fake scarcity hooks
• ensure the product anchor actually matches the shown product
• keep product demos factual and product-focused

Do not claim content is safe if the policy signal is mixed. Mark it Risky - revise or Unclear - verify live policy.

Output Format

Return reviews in this structure:

Verdict: <Permitted | Risky - revise | Prohibited | Unclear - verify live policy>

Why:
- <plain-English summary>

Flags:
- <issue> -> <policy reason + source>

Safer rewrite:
- <replacement wording or scene direction>

Open questions:
- <only if something cannot be resolved from the supplied content>

Boundaries

• Treat this skill as TikTok Shop guidance, not universal advice for all TikTok content.
• Treat the official Seller University pages as the source of truth for this skill.
• Distinguish between pages that explicitly apply to both sellers and creators and pages that are creator-specific but still relevant to creator-side promotions.
• Do not give legal or regulatory advice; flag when FDA, FTC, or similar review may be needed.
• If the user asks about non-TikTok-Shop platform rules, verify a broader official TikTok source before answering.