Password strength analyzer with entropy calculation and crack time estimation.
Single tool with clear naming (security_check_password) and adequate description (194 chars, within baseline 34-392 range). Input schema present with type and description for the password parameter. However, output schema is completely undocumented, the tool returns score, strength rating, crack time, entropy bits, and suggestions, but none of these fields are formally specified. No parameter constraints (min/max length), no error handling guidance, and no documentation of what constitutes weak/fair/strong/excellent ratings. Tool description mentions 'Do NOT use for...' which is helpful for disambiguation but lacks actionable recovery guidance if misused.
Use this when you need to evaluate password strength or security. Returns a score 0-100, strength rating (weak/fair/strong/excellent), estimated crack time, entropy bits, and improvement suggestions. Checks against top 1000 common passwords. Do NOT use for hash generation — use crypto_generate_hash. Do NOT use for input validation — use data_validate_json.
Output schema completely undocumented. Tool returns score (0-100), strength rating (weak/fair/strong/excellent), crack time, entropy bits, and suggestions, but none of these fields are formally specified in the tool definition. LLMs cannot plan downstream operations or extract structured data without knowing the response shape.
No input validation constraints. Password parameter lacks minimum/maximum length, character restrictions, or encoding requirements. Tool accepts any string, but logic assumes UTF-8 and may fail on edge cases (empty string, very long input, special encodings).
No error handling guidance. If password is null, empty, or causes an exception, the tool returns a generic 400 error with 'Missing required field: password'. No recovery hints, no guidance on what constitutes valid input, no actionable next steps for the LLM.
Inferred effective spec: 2025-06-18+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | C | 62 | 2025-06-18+ | v2 |
Strength rating thresholds not documented. Description mentions 'weak/fair/strong/excellent' but does not specify the score ranges that map to each rating. LLMs cannot reliably interpret results without knowing: score 0-25 = weak, 26-50 = fair, etc.
No idempotency guarantee documented. Tool description does not state whether repeated calls with the same password always return identical results. Agents may retry on ambiguous failures, if results vary, duplicate side effects could occur (though this tool is read-only, the pattern should still be explicit).