customize
Pass
Audited by Gen Agent Trust Hub on Aug 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides access to the vendor's administrative CLI (ncl) and a script (scripts/q.ts) that allows the agent to execute arbitrary SQL queries against the system's central and session databases.\n- [COMMAND_EXECUTION]: It enables the configuration of new MCP (Model Context Protocol) servers using arbitrary shell commands via the 'ncl groups config add-mcp-server --command' pattern.\n- [COMMAND_EXECUTION]: The agent is instructed to manage system-level persistence and service lifecycles using 'launchctl' on macOS or 'systemctl' on Linux.\n- [REMOTE_CODE_EXECUTION]: The skill documents an automated workflow for installing new channel adapters that involves fetching code from remote git branches and executing local build processes.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it performs high-privilege system modifications based on user-provided descriptions of desired changes.\n
- Ingestion points: Interactive user requests for modifying behavior, adding channels, or creating integrations.\n
- Boundary markers: None identified; instructions do not include delimiters or warnings to ignore embedded instructions in user requests.\n
- Capability inventory: Wide-ranging permissions including file system writes (CLAUDE.md, .env), database manipulation (SQL), and shell command execution (ncl, service managers).\n
- Sanitization: No evidence of input validation or sanitization before passing user intent into administrative commands or configuration files.
Audit Metadata