The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the nvfleetint CLI tool to perform investigations, retrieve node statuses, and collect alert history. These operations are restricted to data collection and are appropriate for the skill's stated purpose of incident analysis.
[EXTERNAL_DOWNLOADS]: The documentation provides a link to the official NVIDIA GitHub repository for downloading the required client binaries. This reference targets a trusted vendor resource.
[DATA_EXFILTRATION]: While the skill performs web searches for corrective action research, it includes strict instructions to exclude sensitive data such as hostnames, node UUIDs, and tenant information from search queries to prevent data leaks.
[PROMPT_INJECTION]: The skill manages the risk of indirect prompt injection through systematic data handling:
Ingestion points: Data enters the agent's context from nvfleetint command outputs and web search results.
Boundary markers: The workflow requires response validation against a predefined CLI contract and verification of report structure.
Capability inventory: The skill involves shell command execution and file writing to a workspace.
Sanitization: The html-theme.md reference mandates HTML-escaping of all dynamic values, and the workspace workflow ensures proper filename sanitization and secure temporary directory handling (umask 077).

node-rca-rcca