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。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/majiayu000/claude-skill-registry --skill idea-genie命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
LLM token logprobs and calibration. Per-decision confidence, ECE, Brier, reliability diagrams, low-confidence triage.
Analyze LLM token logprobs and calibration. Use for per-decision confidence, ECE, Brier scores, reliability diagrams, and low-confidence triage.
回顾最近 N 天的 Claude Code 使用记录——扫描原始会话数据,按主题分组汇总"我都做了什么",并从个人操作系统视角输出模式、风险与增删建议。当用户说 /recap、"看看我这几天做了什么"、"回顾一下我最近的会话"、"这两天我用 claude 干了啥"、"活动回顾" 时使用。
基于 SOC 职业分类
正在显示 SKILL.md
| name | idea-genie |
| description | Generate an evidence-grounded opportunity |
Turn an open-ended question into a small, evidenced opportunity portfolio. This skill explores; it does not select work, create beads, or write Discovery's BDD intent packet.
research when the repository facts are not already available.assumptions; never blend them into observations.idea-portfolio.v1, run scripts/validate-output.sh, and hand the
artifact to discovery. Discovery alone shapes and persists executable BDD
intent.If every candidate overlaps existing work or lacks support, emit
status: no-new-work, an empty candidate list, and observations showing why.
An honest empty portfolio is a successful result.
Required fields are schema_version, status, observations, assumptions,
candidates, and termination. Candidate portfolios require cited evidence,
an explicit overlaps array, and a complete scenario. Termination records the
reason and novel_candidates_last_pass: 0.
The validator is the machine boundary:
skills/idea-genie/scripts/validate-output.sh <portfolio.json>
Executable behavior: references/idea-genie.feature.