Loading ToolBench
Convert natural language like 'every Monday at 9am' into valid cron expressions with explanation and next run times.
Single tool 'utility_generate_crontab' has a well-structured input schema with proper type definitions and a clear description. However, the description is verbose (400+ chars) and includes example output that may cause LLM literal reuse. Output schema is documented in prose but not as a formal JSON Schema. Parameter 'description' has a good description (72 chars). The tool lacks error handling guidance, confidence scoring explanation, and does not declare idempotency or destructiveness hints. Tool name uses 'utility_' prefix which is generic; 'generate_crontab' alone would be clearer. No validation rules stated for the input description parameter.
Use this when you need to generate a cron expression from a natural language description like "every Monday at 9am" or "every 5 minutes". Returns the cron expression with validation. 1. expression -- the generated cron expression (e.g. "0 9 * * 1") 2. description -- human-readable confirmation of the schedule 3. nextRuns -- next 5 execution timestamps in UTC 4. confidence -- how confident the parser is in the interpretation (high, medium, low) Example output: {"expression":"0 9 * * 1","description":"Every Monday at 9:00 AM","nextRuns":["2026-04-14T09:00:00Z","2026-04-21T09:00:00Z"],"confidence":"high"} Use this FOR creating new scheduled tasks from user instructions, building crontab configs from plain English, or translating business requirements into cron syntax. Use this BEFORE writing cron entries manually. Do NOT use for parsing existing cron expressions -- use schedule_parse_cron instead. Do NOT use for JSON validation -- use data_validate_json instead. Do NOT use for text analysis -- use text_count_words instead.
Output schema documented in prose only; no formal JSON Schema provided. LLMs cannot parse unstructured output descriptions reliably.
Tool description includes example output JSON. LLMs tend to reuse example values literally rather than adapting to context, risking hallucinated cron expressions.
No error handling guidance. What happens if the natural language is ambiguous? How should the LLM interpret low confidence? No recovery path documented.
Tool name 'utility_generate_crontab' uses generic 'utility_' prefix. Clearer name: 'generate_crontab' or 'parse_schedule_to_cron'.
No tool annotations (idempotentHint, readOnlyHint). Tool is read-only and idempotent but this is not declared in the schema.
Inferred effective spec: 2025-06-18+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 61 | 2025-06-18+ | v2 |