The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests untrusted data from the web via search results and web_fetch to curate newsletter content, which presents an indirect prompt injection surface.
Ingestion points: External news articles and web content retrieved via search tools and web_fetch in Phase 2.
Boundary markers: The instructions do not specify explicit delimiters or warnings to ignore instructions embedded within the fetched content.
Capability inventory: The skill utilizes web_fetch for reading external content and has the capability to write the generated newsletter to the /mnt/user-data/outputs/ directory.
Sanitization: No specific sanitization or filtering steps are defined for the fetched external content before it is processed by the agent.
[SAFE]: No malicious patterns, obfuscation, or unauthorized access attempts were detected. The skill follows legitimate coding and documentation standards for its stated purpose.

newsletter-generation