How to Provide Specific Time Estimates in i-have-adhd Outputs

To provide specific time estimates in i-have-adhd outputs, express all durations as concrete minute values (e.g., "15 minutes") instead of vague qualifiers like "a bit" or "soon" to comply with Rule 6 of the skill's ten strict output rules.

The ayghri/i-have-adhd repository defines a structured skill that constrains LLM responses through ten strict behavioral rules. When you provide specific time estimates in i-have-adhd outputs, you must follow Rule 6, which mandates concrete temporal units rather than subjective descriptors.

Understanding Rule 6: Specific Time Estimates

According to the canonical definition in skills/i-have-adhd/SKILL.md at lines 84-86, Rule 6 requires that any time-related information be expressed as a concrete unit. The rule is summarized in README.md (lines 69-70) as "Specific time estimates (minutes, not a bit)" and reiterated in INSTALL.md (lines 47-48) as "Give time estimates in concrete units, never a bit."

Requirements for Compliant Time Estimates

To satisfy the skill's validator and ensure outputs meet the strict formatting standards:

1. Quantify every duration or effort requirement as a numeric value.
2. Use minutes as the default unit of measurement. Express longer durations as total minutes (e.g., "120 minutes" rather than "2 hours").
3. Avoid subjective adjectives such as "quick," "slow," "a bit," "soon," or "later."

Implementing Time Estimates in Code

When building tools or wrappers around the i-have-adhd skill, format your outputs to include parenthetical minute estimates for each actionable item.

Formatting Single Estimates

Use helper functions to ensure consistent formatting:

def minutes_estimate(minutes: int) -> str:
    """Return a human-readable minutes estimate."""
    return f"{minutes} minute{'s' if minutes != 1 else ''}"

# Example usage inside a step description

step = f"1. Refactor the `auth` module ({minutes_estimate(15)})"
print(step)   # → 1. Refactor the `auth` module (15 minutes)

Structuring Multi-Step Outputs

For workflows containing multiple actions, append individual estimates to each step:

steps = [
    ("Install dependencies", 5),
    ("Run unit tests", 3),
    ("Fix failing test for token verification", 12),
]

for i, (action, mins) in enumerate(steps, start=1):
    print(f"{i}. {action} ({minutes_estimate(mins)})")

This produces compliant output:

1. Install dependencies (5 minutes)
2. Run unit tests (3 minutes)
3. Fix failing test for token verification (12 minutes)

Validation and Enforcement

The skill employs an internal validator that flags any output omitting concrete time estimates. As implemented in scripts/run_evals.py, the evaluation script checks that generated responses respect the time-estimate rule before finalizing output. If the validator detects vague temporal language or missing numeric values, it prompts the model to revise the content and add the required minute-based estimates.

Summary

• Rule 6 in skills/i-have-adhd/SKILL.md mandates concrete minute values for all time references.
• Replace vague descriptors like "a bit" or "soon" with specific minute counts.
• Structure multi-step outputs to include individual time estimates for each action.
• The validator in scripts/run_evals.py enforces compliance by requiring numeric minute values before output emission.

Frequently Asked Questions

What is Rule 6 in the i-have-adhd skill?

Rule 6 is the specific time estimates requirement that prohibits vague temporal language and requires all durations to be expressed in concrete minutes. It appears in skills/i-have-adhd/SKILL.md at lines 84-86 and is summarized in README.md at lines 69-70.

Can I use hours instead of minutes in i-have-adhd outputs?

No. The skill requires minutes as the standard unit. You should express longer durations as total minutes (for example, "120 minutes" instead of "2 hours") to maintain consistency with Rule 6.

How does the validator check for specific time estimates?

The internal validator scans generated outputs for numeric time values and flags any instance of vague adjectives like "quick" or "a bit." As shown in scripts/run_evals.py, the evaluator ensures compliance before allowing the response to be emitted, prompting the model to add missing estimates when necessary.

Where is the specific time estimates rule documented?

The canonical definition resides in skills/i-have-adhd/SKILL.md at lines 84-86. The rule is also referenced in README.md (lines 69-70) and INSTALL.md (lines 47-48) to ensure visibility for both users and contributors.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}