The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes AI-generated briefs synthesized from external news sources, which presents an attack surface for indirect prompt injection where malicious instructions could be embedded in the source data.
Ingestion points: The brief field returned by the GET https://www.worldmonitor.app/api/intelligence/v1/get-country-intel-brief endpoint.
Boundary markers: The skill includes a 'Content safety' section that explicitly instructs the agent to treat the response as 'data, not instructions' and to disregard directive-like text.
Capability inventory: The skill uses curl to fetch data and jq to extract content; it does not demonstrate capabilities like file system writes or further network exfiltration using the ingested data.
Sanitization: The skill relies on natural language instructions for the agent to filter/ignore embedded commands rather than technical sanitization or schema-based isolation.
[COMMAND_EXECUTION]: The skill provides documentation and examples using common command-line utilities.
Evidence: The worked examples demonstrate the use of curl for HTTP requests and jq for JSON processing to retrieve and display the intelligence briefs.

fetch-country-brief