Loading ToolBench
MCP server for the Slovak Collection of Laws (Zbierka zakonov) via static.slov-lex.sk - Slovak legislation with national ELI identifiers and verifiable citations.
Four tools with complete schemas, clear descriptions, and proper error handling. Tool names follow verb_noun pattern (sk_get_*). All parameters have types and descriptions. Schemas are well-structured with examples. Error handling includes structured codes (invalid_arg, not_found, upstream_error) with recovery guidance. Main gaps: descriptions are domain-specific but could be more LLM-optimized for general agents; no pagination support despite potential for large text responses; limited parameter validation hints in descriptions.
Declare what this connector covers, how it is sourced, and what it does NOT cover.
Fetch Slovak act metadata by year and number.
Fetch the full text of a Slovak act (a consolidated version).
List the consolidated versions of a Slovak act.
sk_coverage tool name lacks action verb clarity. 'coverage' is a noun; 'describe_coverage' or 'get_coverage_info' would better signal intent to LLMs.
Parameter descriptions lack validation constraints. E.g., year/number accept ranges (1918 - 2100, >0) but constraints are only in code comments, not in parameter descriptions where LLMs read them.
sk_get_text returns full act text without pagination or length limits. Large acts could exceed context windows. No limit documented in description.
version_id parameter in sk_get_text accepts free-form strings ('20240701' or 'vyhlasene_znenie'). Should document exact format constraints and valid enum values if known.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 65 | 2026-07-28+ | v2 |
Error handling returns structured codes but does not include the invalid value in error messages. E.g., 'year=2101 is out of range' is better than generic 'year is out of range'.