The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The script scripts/fetch_xhs_trends.py connects to an external API at https://onetotenvip.com/skill/cozeSkill/getXhsCozeSkillData. This domain is an unverified third-party service with no clear association with the vendor.
[EXTERNAL_DOWNLOADS]: The Python script scripts/fetch_xhs_trends.py explicitly disables SSL/TLS certificate verification by setting context.verify_mode = ssl.CERT_NONE. This is a critical security flaw that exposes the data transmission to Man-in-the-Middle (MITM) attacks, potentially allowing attackers to intercept or modify the fetched data.
[COMMAND_EXECUTION]: The skill's operation relies on executing the bundled Python script scripts/fetch_xhs_trends.py with user-controlled arguments, which is defined as a mandatory step in the workflow instructions in references/core_workflow.md.
[DATA_EXFILTRATION]: User-provided search keywords are transmitted to the external API at onetotenvip.com. This involves sharing user input and interests with a third-party service, which users should be aware of.
[PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection as it ingests and processes content from an external third-party source without sanitization or boundary markers.
Ingestion points: External content is retrieved from onetotenvip.com via scripts/fetch_xhs_trends.py and used for content analysis and generation.
Boundary markers: No markers or delimiters are present to distinguish untrusted external content from internal instructions.
Capability inventory: The agent can execute shell commands (running the script) and perform web searches using web_search.
Sanitization: No sanitization or validation is performed on the titles or descriptions fetched from the external API before they are processed by the model.

xhs-copywriter