cml-node-operations

Pass

Audited by Gen Agent Trust Hub on Mar 6, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSCREDENTIALS_UNSAFEPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill provides the execute_command tool, which allows for the execution of CLI commands on network nodes within a CML environment using pyATS.\n- [EXTERNAL_DOWNLOADS]: The skill documentation indicates a dependency on the cml-mcp Python package, which must be installed via pip.\n- [CREDENTIALS_UNSAFE]: The skill relies on environment variables (CML_USERNAME, CML_PASSWORD) for authentication, which is a standard but sensitive method for managing server access.\n- [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its processing of untrusted data from network devices.\n
  • Ingestion points: External data enters the agent context through get_node_console_log (device logs), execute_command (CLI output), and get_node_config (configurations) as described in SKILL.md.\n
  • Boundary markers: No explicit delimiters or instructions are used to signal the agent to ignore or isolate instructions that might be embedded in the retrieved device data.\n
  • Capability inventory: Across its tools, the skill can execute commands (execute_command), modify node configurations (set_node_config), and perform lifecycle actions (stop_node, wipe_node) via the CML API.\n
  • Sanitization: There is no evidence of data sanitization or validation performed on the information retrieved from nodes before it is interpreted by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 6, 2026, 12:31 AM