The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted data from external sources including TikTok video metadata, SEO keyword suggestions, and AI-answer citations.
Ingestion points: Data returned from UnifAPI seo/*, geo/*, and tiktok/* endpoints described in SKILL.md.
Boundary markers: Absent. The instructions do not specify delimiters or guidelines for the agent to ignore instructions that might be embedded in the external search or social media data.
Capability inventory: The skill performs network operations (API calls) and reads local markdown files.
Sanitization: Absent. There is no mention of filtering or escaping external content before processing it for marketing analysis.
[EXTERNAL_DOWNLOADS]: The skill fetches live marketing data from mcp.unifapi.com. This is a vendor-owned resource belonging to the skill author ('unifapi-agent').
[DATA_EXFILTRATION]: The skill accesses local project files (e.g., .agents/product-marketing.md) to establish context for the venue. While this data is used to inform research queries, it is processed and potentially shared with the external UnifAPI service during the search for cuisine trends.

menu-demand-radar