The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to download the community MCP server from the CiscoDevNet repository on GitHub.
Evidence: git clone https://github.com/CiscoDevNet/thousandeyes-mcp-community.
[COMMAND_EXECUTION]: The installation process for the community server involves executing shell commands to clone the repository and install dependencies.
Evidence: pip install -r requirements.txt and npx mcp-remote.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests and processes untrusted data from network test results, BGP routing information, and dashboard metrics from the ThousandEyes API.
Ingestion points: Results from tools like te_get_test_results, te_get_path_vis, and te_get_dashboard_widget (SKILL.md).
Boundary markers: None explicitly defined in the prompt instructions to isolate external data from agent instructions.
Capability inventory: The skill primarily performs read-only operations but includes the capability to trigger 'Instant Tests' (active network probes) through the official server.
Sanitization: No explicit sanitization or validation logic for external data is described in the skill instructions.

te-network-monitoring