The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill enables the generation and execution of arbitrary code within isolated environments via tools such as mcp__flow-nexus__sandbox_execute. This is a core feature of the platform for testing and running applications.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data that could potentially contain malicious instructions.
Ingestion points: Data enters the system through tools like mcp__flow-nexus__challenge_submit (solution code), mcp__flow-nexus__app_store_publish_app (application source code), and mcp__flow-nexus__seraphina_chat (user messages).
Boundary markers: There are no specific delimiters or warnings documented in the tools to help the AI distinguish between data and instructions.
Capability inventory: The skill possesses capabilities to write files (mcp__flow-nexus__sandbox_upload), execute code (mcp__flow-nexus__sandbox_execute), and deploy applications (mcp__flow-nexus__template_deploy).
Sanitization: The skill does not explicitly mention sanitization or validation logic for the content it processes.
[COMMAND_EXECUTION]: The skill allows the execution of shell commands and script installation within sandbox environments using the startup_script and run_commands parameters in the sandbox_create and sandbox_configure tools.

flow-nexus-platform