菜单
Spawn N independent agents that each invoke the write-theory skill, producing N distinct root theories
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
Design and propose the next step: either a regular data-gathering experiment, literature search, or a concrete solution candidate.
Determine the type of proposal given, execute/delegate accordingly, and return the appropriate resulting ID (experiment ID, literature search ID, or solution ID).
Score and rank solution candidates relative to the research goal, and update parent theory scores.
Integrate recent interpretations from the interpretation log into the associated theory.
Rank the given experiments based on their importance for evaluating theories.
Interpret the results of newly run experiments, literature searches, or solution candidates, and append the findings as new sections to the interpretation log inside the theory folder.
| name | write-n-theories |
| description | Spawn N independent agents that each invoke the write-theory skill, producing N distinct root theories |
| argument-hint | N, exploration ID (E_...), optional literature ID (L_...), and the phenomenon to explain |
You are the N-Theories Coordinator. Your only job is to fan out write-theory across N agents so the caller ends up with N distinct root theories. You do not write any theory yourself.
Arguments: $ARGUMENTS
Parse:
N — number of theories to generate (required).<EXPLORATION_ID> — required, like E_20260414_....<LITERATURE_ID> — optional, like L_20260414_....<PHENOMENON> — free-form description of the phenomenon to explain (required).If N or the exploration ID or the phenomenon is missing, abort with a clear error.
swarm that instructs each swarm agent to invoke the write-theory skill with the exact arguments above (exploration ID, optional literature ID, phenomenon) and return only the resulting T_... theory ID as its final message. Tell each agent that divergence between theories is expected and desired — they should not coordinate — but they must all explain the same phenomenon from the same inputs./swarm "<task>" N=<N>.T_... ID from its report. If any swarm agent failed to produce a T_... ID, note it in your output but do not retry (the caller decides whether to re-run).