The Agent Skills Directory

[SAFE]: The skill consists entirely of documentation and usage instructions for UI components in a documentation platform. No malicious code, exfiltration patterns, or injection attempts were detected.
[INDIRECT_PROMPT_INJECTION]: The skill describes components like <LinkEmbed /> which are designed to ingest metadata from external URLs. This constitutes a potential attack surface for indirect prompt injection if the agent processes malicious metadata.
Ingestion points: External metadata fetched by <LinkEmbed /> (title, description, images).
Boundary markers: None explicitly defined in the documentation for handling external metadata.
Capability inventory: No risky capabilities (subprocess, file-write, network ops) are utilized by this skill's instructions.
Sanitization: Not addressed in the documentation; however, within the scope of this skill's static instructions, no exploitable vulnerability is introduced, and the functionality is consistent with its stated purpose as a documentation tool.

blocks