The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches platform inventories, indices, and skill definitions from ai.syncfusion.com and github.com/syncfusion. These are official vendor resources necessary for the skill's routing and configuration logic.
[COMMAND_EXECUTION]: Uses standard package managers (npm, dotnet, flutter) to install Syncfusion libraries and official agent skills via the skills CLI. It also utilizes npx to execute vendor-provided Model Context Protocol (MCP) servers for documentation retrieval.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from remote inventory files to guide component selection. While this presents an entry point for remote content to influence agent behavior, it is an integral part of the vendor's onboarding workflow.
Ingestion points: Remote files such as https://ai.syncfusion.com/<platform-slug>/inventory.txt and llms.txt.
Boundary markers: Absent; the skill relies on the agent to interpret the retrieved inventory data for component routing.
Capability inventory: Includes command execution (npm, npx, dotnet, skills), file modification (project manifests), and network access.
Sanitization: None performed on the remote inventory text; the agent is expected to match requirements against descriptions.
[DYNAMIC_EXECUTION]: Dynamically downloads and runs official MCP server packages using npx -y and dotnet tool run. These tools are used to supplement the agent's knowledge with current vendor documentation.
[CREDENTIALS_UNSAFE]: References API and License keys throughout the licensing and MCP setup documentation. The skill provides explicit safety instructions to prevent these credentials from being hardcoded, committed to version control, or exposed in logs and chat outputs, aligning with security best practices for secret management.

agent-onboarding