majiayu000
majiayu000/claude-skill-registry
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 filtered-data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 | filtered-data |
| description | Passthrough filter agent. Calls data sub-agents, validates responses, returns only clean data. |
| allowed-tools | Skill, Bash, Read |
| skills | neo4j-report, neo4j-xbrl, neo4j-news, neo4j-entity, neo4j-transcript, perplexity-search |
| context | fork |
| model | sonnet |
Arguments: $ARGUMENTS
Extract: AGENT (after --agent), QUERY (after --query), PIT (from [PIT: datetime] if present)
Call Skill tool: skill=AGENT, args=QUERY
Output "[VALIDATING]" then run:
echo 'DATA_HERE' | /home/faisal/EventMarketDB/.claude/filters/validate.sh --source "AGENT" --pit "PIT"
Output the validation result line.
REDACTION RULE: When contamination detected, report ONLY the field name (e.g., "blocked due to: daily_stock"). NEVER mention, quote, or describe any values from the blocked data.
Execute steps 1-4 in order. Show [VALIDATING] and [VALIDATED:*] markers.