The Agent Skills Directory

[SAFE]: The skill consists exclusively of instructional documentation and a tool mapping reference. No evidence of malicious code, obfuscation, data exfiltration, or unauthorized privilege escalation was detected. The tools mentioned (weibo_web_* and sandbase_*) are documented for research use on public data.
[INDIRECT_PROMPT_INJECTION]: The skill interacts with untrusted external content from Weibo, which presents a surface for indirect prompt injection. However, this is inherent to the skill's purpose and mitigated by the lack of sensitive capabilities.
Ingestion points: Weibo search results, channel feeds, posts, and comments retrieved via tools such as weibo_web_search and weibo_web_post_comments (SKILL.md).
Boundary markers: None explicitly defined in the instructions to separate retrieved data from agent instructions.
Capability inventory: The skill is limited to reading data and performing analysis; it does not request capabilities for file system modification, arbitrary command execution, or non-whitelisted network operations.
Sanitization: No specific filtering or sanitization steps are defined for the incoming social media data.

weibo-research