The Agent Skills Directory

[COMMAND_EXECUTION]: The skill facilitates the execution of a local Python script scripts/lint_rsc.py to perform heuristic analysis on RouterOS scripts. The script is benign, relying on standard Python libraries to perform regex-based static analysis without network access or sensitive file interaction.
[INDIRECT_PROMPT_INJECTION]: The skill operates on user-provided RouterOS configuration files, which presents a surface for potential indirect prompt injection if those files contain malicious instructions hidden in comments.
Ingestion points: RouterOS .rsc files provided by the user and processed by the agent or the lint_rsc.py script.
Boundary markers: The skill workflow explicitly requires validation via a linter and encourages the use of dry-run and error-handling blocks to isolate execution risks.
Capability inventory: The agent has the capability to write files, execute the local linter via subprocess, and suggest RouterOS CLI commands.
Sanitization: The lint_rsc.py utility implements heuristic checks to flag destructive commands (e.g., system resets), unconditional bulk deletions, and excessive policy grants.
[EXTERNAL_DOWNLOADS]: The documentation and example files (references/EXAMPLES.md) reference fetching remote scripts from https://example.com/config.rsc. This is provided as an educational example for the RouterOS /tool fetch command and targets a reserved documentation domain.

mikrotik-routeros-rsc