The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple local scripts (e.g., $PYATS_MCP_SCRIPT, $ISE_MCP_SCRIPT, $GAIT_MCP_SCRIPT) via an execution wrapper to interact with network infrastructure and security servers.
[EXTERNAL_DOWNLOADS]: Fetches and executes the nvd-cve-mcp-server package from the npm registry using npx to perform vulnerability lookups against the National Vulnerability Database.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it processes untrusted data from network device configurations. • Ingestion points: Running configurations and command outputs are captured via pyats_show_running_config and pyats_run_show_command. • Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore embedded instructions within device configurations. • Capability inventory: The skill possesses the ability to execute shell commands, perform network operations to Cisco ISE, and record audit trails in GAIT. • Sanitization: No explicit sanitization, escaping, or validation of the ingested device configuration data is performed prior to processing.

pyats-security