✓
Purpose & Capability
Name/description focus on repurposing short-video/social content; the declared runtime actions (web_search, web_fetch, images_search_and_download, file_write, optional cloud_deploy) are appropriate and proportional for that purpose. No unrelated env vars, binaries, or install steps are requested.
ℹ
Instruction Scope
SKILL.md stays focused on finding recent (1-5 day) sources, extracting repurposable material, producing platform-specific scripts/titles/tags, and saving outputs. This is coherent. Two items to note: (1) the skill repeatedly authorizes broad web_search/web_fetch and image download operations — which is expected but can fetch copyrighted or sensitive material, and (2) the '定时任务规则' (automatic scheduled generation with '无需人工触发') gives the skill discretion to run recurring work; how that schedule is implemented by the host/platform is unspecified and could lead to unexpected background activity.
✓
Install Mechanism
Instruction-only skill with no install spec and no code files. Lowest-risk install surface; nothing will be written to disk by an installer. All runtime behavior depends on the host platform's tool bindings.
ℹ
Credentials
The skill declares no required environment variables or credentials, which is appropriate. However, it references tools (cloud_deploy, images_search_and_download, web_fetch) that on many platforms require credentials/permissions to operate; the SKILL.md does not state which credentials will be needed or how publishing credentials will be handled. Verify any runtime credential grants before enabling those tools.
!
Persistence & Privilege
The skill sets always:false (good) but the documentation explicitly describes unattended scheduled tasks and automatic generation/publishing. That is a functional mismatch: the skill asks to run recurring, unsupervised work but does not declare a persistent/install mechanism (always:true) or explain how scheduling is authorized. Combined with optional cloud_deploy (which can publish content), this creates a risk of the agent autonomously generating and deploying content if the host platform grants those tool permissions. Confirm how scheduling and deployment are controlled by the host before installing.
What to consider before installing
This skill appears to do what it says (find recent trending items, craft platform-specific short-video scripts/titles/tags, and save outputs). Before installing, check the following:
- Tool permissions: Ask which platform tools will be bound to web_search/web_fetch/images_search_and_download/file_write/cloud_deploy. In particular, limit or review any cloud_deploy/publish permission — do not grant publish credentials unless you explicitly want the agent to post on your behalf.
- Scheduling behavior: The SKILL.md says it will 'automatically' generate content and '无需人工触发'. Confirm whether the host will actually run scheduled jobs and whether you can opt out or review drafts before publication.
- Data and copyright: The skill will fetch images and external materials; ensure you understand copyright/legal risk when reusing third-party content and whether the skill performs adequate provenance checks.
- Fact‑checking and sensitivity: The skill promises fact-checking but does not describe sources or thresholds; if you need strict compliance (legal, medical, political), require manual review steps.
- Provenance and logging: Prefer installations that log actions, show drafts before deployment, and allow you to revoke tool permissions. Because the owner is unknown and there's no homepage, consider running the skill in a constrained environment (no publish credentials, file writes to a sandboxed workspace) until you’re comfortable.
If you want, I can draft a short permission checklist or recommended host-binding policy to safely run this skill (e.g., disable cloud_deploy, require human approval for scheduled runs, sandbox file_write).
