The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill uses the mcp__flow-nexus__sandbox_execute tool to run Python code generated at runtime. Although the provided examples follow a set template for monitoring performance, this capability allows the agent to execute arbitrary scripts within the sandbox environment.
[COMMAND_EXECUTION]: The Python scripts executed via the sandbox utility leverage the psutil library to collect sensitive system information, including CPU load, virtual memory usage, disk I/O counters, and network interface statistics.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external performanceData, systemMetrics, and systemTopology.
Ingestion points: System metrics and performance data are ingested through tools like mcp__sublinear-time-solver__analyzeMatrix and the Python monitoring script.
Boundary markers: No specific delimiters or instructions to ignore embedded commands within the ingested data are present in the examples.
Capability inventory: The skill possesses the ability to execute code via mcp__flow-nexus__sandbox_execute and solve optimization problems using mcp__sublinear-time-solver__solve.
Sanitization: No explicit sanitization or validation logic is shown for the data passed into the optimization or execution tools.

agent-performance-optimizer