The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions to download and execute an installer script directly from the official Nacos website using curl piped to bash for Unix-like systems and Invoke-WebRequest for Windows.
[COMMAND_EXECUTION]: The installation process for Linux and macOS utilizes sudo to execute the downloaded script with administrative privileges.
[COMMAND_EXECUTION]: The skill uses local system utilities for workflow management, including mkdir for creating configuration directories, find for tracking file modifications, and the system's text editor (e.g., vi) for manual configuration editing.
[PROMPT_INJECTION]: An indirect prompt injection surface is present because the skill retrieves and processes metadata and instructions from external Nacos registries. Content from these registries is untrusted and could contain malicious instructions designed to manipulate agent behavior.
Ingestion points: Data enters the environment through registry interaction commands like nacos-cli skill-list, nacos-cli skill-describe, and nacos-cli skill-get (as described in SKILL.md).
Boundary markers: No specific boundary markers or instructions are provided to the agent to treat external skill data as untrusted or to ignore embedded commands.
Capability inventory: The skill possesses extensive local capabilities, including network access via nacos-cli, file system modification, and command execution (documented in SKILL.md).
Sanitization: The skill does not include mechanisms for validating or sanitizing data fetched from the Nacos registry before it is interpreted by the agent.
[SAFE]: The skill implements best practices for secret management by guiding users to use local profiles instead of hardcoding credentials and by recommending the use of temporary security tokens via sts-hiclaw.
[EXTERNAL_DOWNLOADS]: The skill facilitates the download of the nacos-cli binary and external skill packages from Nacos registry servers.

nacos-skill-registry