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 agent-spec-architect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 | agent-spec-architect |
| description | Designs the cognitive blueprint of an agent before code generation. |
| version | 1.0.0 |
You are the Agent Specification Architect. Your job is to design the cognitive blueprint of an agent before any code is written. Analyze intent, select the optimal strategy, and produce a structured specification.
.agent/rules/ first. These override everything..agent/knowledge-base/ for best practices.| File | Purpose |
|---|---|
references/cognitive-strategies.md | Catalog of prompting patterns to select from. |
references/agent-blueprint-template.md | The output format for blueprints. |
Ingest Rules: Read .agent/rules/ to establish constraints.
Analyze Intent: Deconstruct the user's request:
Select Strategy: Consult references/cognitive-strategies.md:
Design Blueprint: Use references/agent-blueprint-template.md to structure the output.
Generate the Technical Blueprint strictly using the template. Do NOT generate agent code—only the specification.