CtrlK
BlogDocsLog inGet started
Tessl Logo

leo-laptop/calculate

Use when the user asks you to calculate, compute, evaluate, or solve a math expression or equation. Triggers on arithmetic, order of operations (PEMDAS), fractions, percentages, exponents, and multi-step math problems.

84

1.00x
Quality

78%

Does it follow best practices?

Impact

94%

1.00x

Average score across 5 eval scenarios

SecuritybySnyk

Passed

No known issues

Overview
Quality
Evals
Security
Files

criteria.jsonevals/scenario-4/

{
  "context": "Tests whether the agent correctly applies left-to-right evaluation for equal-priority operations (multiplication/division and addition/subtraction), avoids the common mistake of evaluating these expressions in the wrong order, and follows the step-by-step presentation format.",
  "type": "weighted_checklist",
  "checklist": [
    {
      "name": "Correct result expr 1",
      "description": "Expression 1 (20 / 4 * 2): correct answer is 10 (not 2.5); agent identifies developer's answer as incorrect",
      "max_score": 10
    },
    {
      "name": "Correct result expr 2",
      "description": "Expression 2 (18 - 5 + 3): correct answer is 16 (not 10); agent identifies developer's answer as incorrect",
      "max_score": 10
    },
    {
      "name": "Correct result expr 3",
      "description": "Expression 3 (100 / 5 / 4): correct answer is 5 (not 80); agent identifies developer's answer as incorrect",
      "max_score": 10
    },
    {
      "name": "Correct result expr 4",
      "description": "Expression 4 (7 * 2 / 4 * 3): correct answer is 10.5; agent identifies developer's answer as correct",
      "max_score": 10
    },
    {
      "name": "Left-to-right rule stated",
      "description": "The audit report explicitly states or demonstrates that equal-priority operations are evaluated left to right",
      "max_score": 15
    },
    {
      "name": "Step-by-step work shown",
      "description": "Each expression includes intermediate results for each operation, not just the final answer",
      "max_score": 10
    },
    {
      "name": "Expression restated",
      "description": "Each expression is restated or quoted clearly before its solution",
      "max_score": 7
    },
    {
      "name": "PEMDAS steps labeled",
      "description": "At least two expressions use labeled PEMDAS step headings (e.g., 'Step 1 — Mult/Div' or equivalent)",
      "max_score": 10
    },
    {
      "name": "Final answer clearly stated",
      "description": "Each expression ends with a clearly labeled correct answer",
      "max_score": 8
    },
    {
      "name": "No right-to-left error",
      "description": "Expression 1 solution does NOT show 4 * 2 = 8 first and then 20 / 8 (which would give 2.5) — division is performed before multiplication because it appears first left to right",
      "max_score": 10
    }
  ]
}

evals

SKILL.md

tile.json