with one click
design-shi-knowledge-base-ops
设计虱知识库运维操作 — raw文章批量入库、来源页/概念页/实体页创建规范、自检脚本
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Menu
设计虱知识库运维操作 — raw文章批量入库、来源页/概念页/实体页创建规范、自检脚本
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
高德地图综合服务,支持POI搜索、路径规划、旅游规划、周边搜索和热力图数据可视化
Knowledge comic creator supporting multiple art styles and tones. Creates original educational comics with detailed panel layouts and sequential image generation. Use when user asks to create "知识漫画", "教育漫画", "biography comic", "tutorial comic", or "Logicomix-style comic".
Generate professional infographics with 21 layout types and 21 visual styles. Analyzes content, recommends layout×style combinations, and generates publication-ready infographics. Use when user asks to create "infographic", "visual summary", "信息图", "可视化", or "高密度信息大图".
Darwin Skill (达尔文.skill): autonomous skill optimizer inspired by Karpathy's autoresearch. Evaluates SKILL.md files using an 8-dimension rubric (structure + effectiveness), runs hill-climbing with git version control, validates improvements through test prompts, and generates visual result cards. Use when user mentions "优化skill", "skill评分", "自动优化", "auto optimize", "skill质量检查", "达尔文", "darwin", "帮我改改skill", "skill怎么样", "提升skill质量", "skill review", "skill打分".
从 DESIGN.md 生成预览图并截图的工作流 - 解决 Playwright 浏览器路径问题和 YAML 解析
Cron jobs auto-deliver final responses to configured targets — send_message to the same target gets deduplicated and skipped. Print report content directly as final response instead.
| name | design-shi-knowledge-base-ops |
| description | 设计虱知识库运维操作 — raw文章批量入库、来源页/概念页/实体页创建规范、自检脚本 |
| tags | ["knowledge-base","design-shi","ingestion"] |
| created | "2026-05-14T00:00:00.000Z" |
| notes | 本 skill 包含的表格是为阅读友好设计的,在飞书等不支持 markdown 表格的平台,请用代码块格式重述表格内容。 |
wiki/
├── 实体/ # 有形实体(人/产品/工具)
├── 概念/ # 方法论/流程/概念
├── 来源/ # 文章来源摘要
├── 对比/ # 对比分析
├── 洞察/ # 定期洞察报告
├── index.md # 内容索引(Agent维护)
└── log.md # 操作日志(仅追加)
| 知识库 | 维护者 | 操作 |
|---|---|---|
| IMA (ima.qq.com) | 其他智能体(自动抓取/整理) | 只读,定期读取已整理的笔记 |
| 本地知识库 (/opt/data/workspace/设计虱-知识库/) | 我(Hermes) | 写入,包含来源页/概念页/实体页 |
关键规则:
ls -lt /opt/data/workspace/设计虱-知识库/raw/ | head -30
检查已入库记录(log.md 末尾),确认文章是否已入库。
已处理文章重命名为 .processed 后缀。
实体页(人/产品/工具/服务)
type: entity、name、tags、sources、created、confidence、memory_layer、relates_toSenseNova.md、绿联DX4600.md概念页(方法论/流程/思维模型)
type: concept、name、tags、sources、created、confidence、memory_layer、relates_to横纵分析法.md、多平台内容分发.md来源页(文章来源摘要)
type: source、name、tags、sources、created、author、confidence、relates_to来源-{文章标题摘要}.md概念/OpenClaw关闭审批.mdindex.md 同步:新增条目追加到对应分类下
log.md 同步:追加 ## YYYY-MM-DD 章节,记录入库文件列表(用 [[文件名]] 格式)
import os, re, yaml
base = "/opt/data/workspace/设计虱-知识库/wiki"
issues = []
for root, dirs, files in os.walk(base):
for f in files:
if f.endswith(".md") and f not in ("index.md", "log.md"):
path = os.path.join(root, f)
with open(path, encoding="utf-8") as fh:
content = fh.read()
fm_match = re.match(r'^---\n(.*?)\n---', content, re.DOTALL)
if not fm_match:
issues.append(f"❌ 缺少frontmatter: {path}")
continue
try:
fm = yaml.safe_load(fm_match.group(1))
except yaml.YAMLError as e:
issues.append(f"❌ YAML错误: {path} — {e}")
continue
if not issues:
print("✅ 自检通过")
else:
for i in issues:
print(i)
注意:relates_to 存的是语义名称(如 "RAG"、"LLM Wiki"),不是文件路径,检查脚本用文件路径匹配会误报。这是知识库设计预期用法,不算问题。
mv "原文件名.md" "原文件名.md.processed"
# 实体页
---
type: entity
name: 实体名
tags: [标签1, 标签2]
sources: [./来源/来源-xxx.md, https://...]
created: 2026-05-14
confidence: 0.8
memory_layer: semantic # semantic | episodic | procedural
relates_to:
- [implements, [[实体B]]]
- [relates_to, [[概念C]]]
---
# 概念页
---
type: concept
name: 概念名
tags: [标签1, 标签2]
sources: [./来源/来源-xxx.md, ...]
created: 2026-05-14
confidence: 0.8
memory_layer: procedural # semantic | episodic | procedural
relates_to:
- [invented_by, [[人名]]]
- [implements, [[工具名]]]
---
# 来源页
---
type: source
name: 文章标题
tags: [标签1, 标签2]
sources: [https://原始链接]
created: 2026-05-14
author: 作者名
confidence: 0.8
relates_to:
- [implements, [[概念名]]]
- [relates_to, [[实体名]]]
---
[implements, [概念]] — 来源实现了某个方法/概念[invented_by, [人名]] — 由某人发明[relates_to, [实体/概念]] — 相关[supersedes, [旧实体]] — 取代了某个旧实体知识库 wiki/ 下的页面已全量向量化,支持语义检索(与 FTS5 关键词检索互补)。
/opt/data/scripts/wiki_embed.py/opt/data/state.db 的 wiki_embeddings 表herald/dmeta-embedding-zh(768维,195MB,通过 Ollama 本地运行)wiki-embed-daily,每天凌晨 04:00 增量执行/opt/data/scripts/wiki_embed.log# 增量模式(只处理新增/修改的文件,按 mtime 判断)
python3 /opt/data/scripts/wiki_embed.py
# 全量模式(重新向量化所有文件)
python3 /opt/data/scripts/wiki_embed.py --full
background=true + notify_on_complete=truewait 工具最大 clamped 到 60s,长任务必须用 tail -f 或 tail -20 检查日志进度@@ 残留:用 patch 的 patch 模式追加 log.md 条目时,多次出现 diff 标记 @@ 被写入文件的 bug。修复方法:log.md 追加改用 Python str.replace() 或 read_file + write_file 全量写入。index.md 的 patch 操作正常,不受影响。/api/embeddings + prompt 参数,不支持 /v1/embeddings + input