The Agent Skills Directory

[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface (Category 8c) because it processes data from X (Twitter) bios and profiles, which are attacker-controlled external sources. Malicious instructions placed in a bio could theoretically attempt to influence the agent's logic during the filtering or enrichment steps.
Ingestion points: External data enters the context via GET /extractions/{id}/results and GET /x/users/{username}.
Boundary markers: The instructions do not define specific delimiters to separate user-controlled profile content from agent instructions.
Capability inventory: The skill is limited to API requests to xquik.com and local data filtering; it does not have file-write or shell execution capabilities.
Sanitization: No explicit sanitization of the bio content is mentioned before it is processed by the agent.
[DATA_EXFILTRATION]: The skill communicates with https://xquik.com/api/v1. As this is the official domain for the skill's vendor (Xquik), these network operations are considered standard functionality for the intended search and extraction service. It uses the XQUIK_API_KEY environment variable for authentication, following secure secret management practices.

find-influencers