From strucmod
跑反事实 / 过渡路径 / 政策实验,对比 baseline。当用户说"反事实"、"counterfactual"、"政策实验"、"transition path"、"如果……会怎样"时调用。
How this skill is triggered — by the user, by Claude, or both
Slash command
/strucmod:counterfactual-runThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
- "把关税从 5% 提到 25% 看 GDP 路径"
| 字段 | 类型 | 说明 |
|---|---|---|
model_name | str | 已求解 baseline |
scenarios | list[dict] | 每个含 name、param_overrides、shock_path |
horizon | int | 默认 200(过渡路径长) |
welfare_metric | enum | cev / lump_sum / none |
output/tables/<model>_cf_<scenario>.csv。verifier 重跑核对。| 文件 | 必产 | 说明 |
|---|---|---|
output/tables/<model>_cf_<scenario>.csv | ✅ | |
output/checkpoints/<model>_cf_<scenario>.json | ✅ | |
output/figures/<model>_cf_<scenario>.{pdf,png} | ✅ |
validated == true 的情况下做反事实verifier| 症状 | 处理 |
|---|---|
| 必备前置未满足 | 报错并指出缺什么;不绕过 |
| 上游产物 stale | 重跑上游 skill;不"凭旧的 checkpoint 继续" |
| 用户口径模糊 | 调用 interview-me skill 反向访谈 |
Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
npx claudepluginhub allblueuk/structural-modeling-for-economists --plugin strucmod