yogsoth-ai/de-anthropocentric-research-engine

select-next-pair

Select the next comparison pair to maximize expected information gain.

View source

How to use

  1. Copy the installation command and review the source package before installing it in your own agent environment.
  2. Check the source document for required tools, credentials and supported hosts.
  3. Describe your input and desired deliverable; review the output against the source material.
npx skills add https://github.com/yogsoth-ai/de-anthropocentric-research-engine --skill select-next-pair

Before you start

  • Installation does not include a model subscription, API allowance or permission to access your files.
  • Repository stars and installation counts describe activity, not an independent quality or security test.
Original skill document

Rendered from the source repository. Headings, examples, code, tables, links, and referenced images are preserved.

select-next-pair

Purpose

Select the next comparison pair to maximize expected information gain while respecting unresolved uncertainty and coverage.

Input contract

yaml
required: [candidate_set, current_ratings, uncertainty_matrix, comparison_budget]
optional: [pair_history, exploration_policy, exclusion_constraints]
constraints: [candidate IDs are stable; expected information gain uses the declared model and available evidence]

Procedure

  1. Enumerate eligible unjudged pairs and compute their current uncertainty and decision relevance.
  2. Estimate expected information gain under the declared comparison model, including the value of resolving a tie or cycle.
  3. Exclude pairs blocked by scope or data constraints, then select the highest-value eligible pair.
  4. Return the selected pair, rationale, alternatives considered, and the budget impact.

Output contract

yaml
produces: [selected_pair, information_gain_estimate, candidate_pair_table, selection_rationale]
delta_fields: [findings, evidence_updates, uncertainties, decisions, recommended_jumps]

Quality gates

  • The selected pair is eligible, not already resolved under the same model, and has a stated information-gain rationale.
  • If all pairs are exhausted or blocked, return an explicit stop reason.

Failure and counterexamples

Do not select a pair solely because it is adjacent in a list. Expected information gain cannot be claimed when uncertainty inputs are absent.

Provenance map

  • resolved: pair-selector

Related by task

from this repository

More skills

All skills