The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill utilizes a remote MCP server hosted at https://mcp.unifapi.com and suggests installing dependencies via npx skills add unifapi-agent/agents. These resources are provided by the skill's author (unifapi-agent) and are required for the skill to fetch live research data.
[COMMAND_EXECUTION]: The documentation includes instructions for setting up the environment using CLI commands such as npx and platform-specific plugin marketplace commands (/plugin marketplace add). These commands facilitate the deployment of the vendor's specialized agents.
[PROMPT_INJECTION]: The skill is designed to ingest and process data from untrusted external sources, creating a surface for indirect prompt injection.
Ingestion points: Untrusted content is retrieved from Reddit comments (reddit/posts/{id}/comments), search engine results (seo/serp), and news searches (news/search) through the UnifAPI service.
Boundary markers: The instructions do not specify the use of delimiters (e.g., XML tags or triple backticks) to isolate external data from the system prompt, nor do they include explicit warnings to ignore instructions embedded within the ingested data.
Capability inventory: The skill is configured as a "read-only" marketing researcher. Its primary capability is reading public data and generating a markdown report; it does not explicitly request file-write or subprocess execution capabilities beyond the initial installation.
Sanitization: No validation or sanitization of the external data is described before it is integrated into the agent's context for analysis.

homeowner-question-content