elicit-weights

SkillDev tools

Lets your agent turn your stated preferences into normalized weights for a set of decision criteria.

Available today. Use it from your connected AI after setup.

Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

Then ask your AI: use the elicit-weights skill

About this skill

Produce a normalized criterion-weight vector using a selected elicitation method.

What this skill tells your AI

The instructions your AI receives, as published by yogsoth-ai/de-anthropocentric-research-engine in skills/elicit-weights/SKILL.md and read by ahel’s review.

Purpose

Produce a normalized criterion-weight vector using a caller-selected elicitation method.

Input contract

required: [criteria, elicitation_method, preference_inputs]
optional: [consistency_threshold, pairwise_scale, missing_preference_policy]
constraints: [weights are nonnegative and sum to 1 within declared tolerance]

Procedure

  1. Validate criteria and method applicability.
  2. Convert preference inputs to a weight vector using the declared method.
  3. Compute consistency diagnostics where the method supports them.
  4. Return weights, diagnostics, and unresolved preference gaps.

Output contract

produces: [weight_vector, consistency_diagnostics, preference_gaps, method_record]
delta_fields: [decisions, uncertainties, open_questions]

Quality gates

  • AHP-style elicitation accepts 2-9 dimensions only.
  • Weight sum is 1.0 within caller-declared tolerance (default +/-0.001).
  • Consistency ratio is reported; CR > 0.1 is flagged when applicable.

Parameterization

Caller supplies criterion schema, method, pairwise/preference scale, tolerance, and consistency policy.

Failure and counterexamples

Reject inapplicable dimensionality, negative weights, missing comparisons, or unreported inconsistency.

Provenance map

  • concept: hypothesis-formation/ahp-weighting
  • concept: convergence/weight-elicitation-sop

Preserved source criteria ledger

sourcecriterion
hypothesis-formation/ahp-weightingAHP applicability range is 2-9 dimensions.
hypothesis-formation/ahp-weightingWeight vector sums to 1.0 within +/-0.001.
hypothesis-formation/ahp-weightingCR > 0.1 is flagged.

Signals

GitHub stars
501
Forks
41
Last commit
Sep 2026
Advanced
Catalog kind
skill
Key
elicit-weights
Source
github.com/yogsoth-ai/de-anthropocentric-research-engine