一键导入
ranking-elo-update
Apply deterministic Elo updates for one completed tournament batch, persist touched hypotheses, and write the ranking update receipt.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Apply deterministic Elo updates for one completed tournament batch, persist touched hypotheses, and write the ranking update receipt.
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Run the full Co-Scientist pipeline for one research run.
Evaluate whether a hypothesis newly entered the current top-k set and update the convergence counter deterministically.
Assemble a dashboard-ready snapshot from the canonical state.
Choose exactly one concrete evolution strategy for the active evolution round.
Run the assumption decomposition and deep verification review for a hypothesis.
Run the iterative evolution loop from the current persisted run state until convergence.
| name | ranking-elo-update |
| description | Apply deterministic Elo updates for one completed tournament batch, persist touched hypotheses, and write the ranking update receipt. |
Goal:
Inputs:
matchesmatchupsstrategyrun_dirk_factortop_k_limitOutputs:
hypotheses/<id>/HYPOTHESIS.json artifacts for all touched hypothesesstate/ranking_update_receipts/<receipt_id>.jsonRankingUpdateReceiptContractContext Loading:
skills/shared-references/schema-index.md.packages/agent_contracts/ranking.py and confirm the exact TournamentMatchContract plus HypothesisMatchupContract shapes.packages/agent_contracts/hypothesis.py and confirm the exact HypothesisContract shape before writing any touched HYPOTHESIS.json artifact.Execution Contract:
from tools import apply_and_persist_elo_updates as the stable invocation surface.packages/run_artifacts/ranking_writeback.py.apply_and_persist_elo_updates(run_dir, matches, matchups, strategy, k_factor=..., top_k_limit=...) -> RankingUpdateReceiptContract.strategy values are placement_tournament and ranked_tournament only.REVIEW/*.json stage artifacts.RankingUpdateReceiptContract.Execution Steps:
skills/shared-references/schema-index.md, then read packages/agent_contracts/ranking.py and packages/agent_contracts/hypothesis.py before writing any touched HYPOTHESIS.json.TournamentMatchContract payloads and the paired matchup payloads as HypothesisMatchupContract.winner_id and that the requested strategy is valid.tools.apply_and_persist_elo_updates(run_dir, matches, matchups, strategy, k_factor=..., top_k_limit=...).RankingUpdateReceiptContract names every completed match ID and touched hypothesis ID.state/ranking_update_receipts/<receipt_id>.json before declaring completion.Artifact Rules:
placement_match_ids for the challenger hypothesis and ranked_match_ids for the defender.ranked_match_ids for both participating hypotheses.tools.apply_and_persist_elo_updates(...) after every Elo batch.tools.sync_hypothesis_review(...); Elo writeback must not clobber those standalone review artifacts.Completion Rule:
hypotheses/<id>/HYPOTHESIS.json artifact has been rewritten in canonical contract form, and the corresponding ranking update receipt exists.