agentsclimarketplace

Hypothesis ranked tournament

Skill panjose/Co-Scientist/skills/hypothesis-ranked-tournament

Scientific agent skills for Claude Code and Codex that turn research goals into auditable hypothesis generation, review, ranking, evolution, and synthesis.

Install
npx -y skills add panjose/Co-Scientist --skill hypothesis-ranked-tournament

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 4 stars4 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Judge one ranked-tournament matchup between two top frontier hypotheses.

SKILL.md

2.0 KB, as published. Nobody here has run it

hypothesis-ranked-tournament

Goal:

  • Judge one ranked-tournament matchup between two top frontier hypotheses.

Inputs:

  • research_plan/RESEARCH_PLAN.json
  • two hypotheses/<id>/HYPOTHESIS.json artifacts with completed review bundles

Outputs:

  • one TournamentMatchContract result for a ranked tournament

Context Loading:

  • Open skills/shared-references/schema-index.md.
  • Read packages/agent_contracts/ranking.py and confirm the exact TournamentMatchContract shape before emitting the ranked result.
  • Read research_plan/RESEARCH_PLAN.json.
  • Read both hypotheses and their review bundles.
  • Confirm that the match strategy is ranked_tournament.

Execution Prompt Contract:

  • System Intent:
    • You are running a deeper comparative judgment for two high-ranked hypotheses whose relative ordering matters.
  • Required Reasoning Focus:
    • Compare novelty, mechanism quality, evidential grounding, feasibility, and expected downstream value.
    • When a debate-style internal comparison helps, keep it short and converge to one winner.
  • Do Not Do:
    • Do not leave the match undecided.
    • Do not ignore review weaknesses that materially change the ordering.
  • Output Shape:
    • Emit one TournamentMatchContract.
    • match_strategy: ranked_tournament
    • winner_id: one of the two hypothesis IDs
    • reasoning: concise but specific
    • debate_turns: optional short internal debate transcript when useful

Execution Steps:

  1. Open skills/shared-references/schema-index.md, then read packages/agent_contracts/ranking.py before emitting the ranked match artifact.
  2. Compare the two frontier hypotheses for ranked ordering.
  3. Choose exactly one winner.
  4. Emit one valid ranked-tournament match result.

Completion Rule:

  • This skill is complete only when one valid ranked match result has been produced with a non-empty winner and reasoning.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.