Install
$ agentstack add skill-vectorpeak-vectorpeak-agent-skills-daily-notes-vp ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ● Network access Used
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
Daily Notes VP
Purpose
Use this skill to append lightweight personal learning records into the raw daily-notes layer.
This skill is for raw capture, not polished Wiki writing. Preserve the user's intent, keep the entry easy to read, and avoid turning a quick note into a heavy template.
Destination
Default destination:
E:\LLM_wiki\LLM_wiki\01.raw\02.DailyNotes\YYYY-MM-DD_YYYY-MM-DD.md
Use 10-day buckets:
- Day 01-10:
YYYY-MM-01_YYYY-MM-10.md - Day 11-20:
YYYY-MM-11_YYYY-MM-20.md - Day 21-end: use the actual last day of the month, such as
YYYY-MM-21_YYYY-MM-28.md,YYYY-MM-21_YYYY-MM-30.md, orYYYY-MM-21_YYYY-MM-31.md
Examples:
E:\LLM_wiki\LLM_wiki\01.raw\02.DailyNotes\2026-06-01_2026-06-10.md
E:\LLM_wiki\LLM_wiki\01.raw\02.DailyNotes\2026-06-11_2026-06-20.md
E:\LLM_wiki\LLM_wiki\01.raw\02.DailyNotes\2026-06-21_2026-06-30.md
If E:\LLM_wiki\LLM_wiki\01.raw\02.DailyNotes does not exist, ask the user before writing somewhere else. Do not silently fall back to the current workspace.
File Header
When creating a new 10-day file, use this header:
---
type: daily-notes
date_range: YYYY-MM-DD_YYYY-MM-DD
created: YYYY-MM-DD
updated: YYYY-MM-DD
tags:
- daily-notes
- raw
- learning
ai-first: true
---
## For future Claude
这是原始学习记录,后续可整理到 Wiki。
When appending to an existing file, update only the updated date. Do not rewrite older entries.
Daily Structure
Each date uses this fixed order:
## YYYY-MM-DD
### Goal-目标
### Question-疑问
### Code-代码
### Concept-概念
### Pitfall-踩坑
If the date section does not exist, create it with all five headings in this order. If a category heading is missing, add it in the correct order.
Categories
Classify each item into exactly one category unless the user explicitly gives multiple items:
Goal-目标: short-term learning goals, small study tasks, today's focus, review plansQuestion-疑问: doubts, questions, interview questions, things the user wants explained laterCode-代码: code questions, code snippets, commands, tiny examples, API usage notes, implementation ideas, or code-shaped material the user wants to revisit later. Treat it as close toQuestion-疑问, but use it when the note is centered on code, command usage, function behavior, configuration snippets, or implementation details.Concept-概念: terms, definitions, methods, models, mechanisms, reusable conceptual notesPitfall-踩坑: debugging lessons, deployment incidents, configuration traps, dependency or route conflicts, "what went wrong / why / how to avoid it next time" experience notes
If classification is uncertain between Question-疑问 and Code-代码, prefer Code-代码 when the title/body contains concrete code, commands, APIs, config, function names, file paths, stack traces, or implementation details; otherwise prefer Question-疑问.
Entry Format
Use a compact entry. Do not include timestamps. Do not add subheadings like "原始记录", "快速整理", or "标签".
#### N. 标题 #tag
正文记录或简短解释。
后续:可选。
Rules:
Nis the next number within the same category for that date.- Put lightweight tags at the end of the title when useful.
- Keep the title as a natural question, goal, or concept name.
- Write in Chinese when the user writes in Chinese.
- Keep the user's original wording when it is concise; lightly polish only for readability.
- Add
后续:...only when there is a real follow-up action.
Example
## 2026-06-17
### Goal-目标
#### 1. 搞清楚 RAG 分块策略 #rag
今天想把 chunk size、chunk overlap、语义边界这几个点串起来。
### Question-疑问
#### 1. 为什么 chunk size 不能固定成一个经验值? #rag #chunking
固定经验值容易切得太小丢上下文,或切得太大降低召回精度。更合理的做法是结合语义边界、文档结构、embedding 表达能力和检索场景来定。
后续:整理成 Wiki/questions 页面。
### Code-代码
#### 1. FastAPI WebSocket 入口如何避免阻塞事件循环? #python #fastapi #async
同步生成器或阻塞检索不能直接在 async handler 里跑,可以用 `asyncio.to_thread(...)` 推进一步,再把事件通过 WebSocket 发回前端。
后续:整理成 FastAPI 异步服务 checklist。
### Concept-概念
#### 1. Semantic Chunking #rag #concept
Semantic chunking 是按语义边界切分文本,而不是机械按 token 数切。它更适合问答和知识库场景,但实现成本更高。
### Pitfall-踩坑
#### 1. FastAPI `/docs` 与项目文档路由冲突 #fastapi #docs #pitfall
FastAPI 默认会把 Swagger UI 挂到 `/docs`。如果项目也想把 MkDocs 或静态课程文档挂到 `/docs`,需要显式把 Swagger 改到 `/api/docs`,否则线上访问 `/docs` 时会被 Swagger 抢占。
后续:部署前把路由约定写进 README 或 guardrail。
Multi-item Input
If the user gives several items at once:
- Split them into separate entries.
- Classify each entry.
- Append them under the same date.
- Number each category independently.
Safety
Do not record:
- API keys, passwords, tokens, credentials, or private identity data
- Large copied documents or source files
- Claims as final truth when the input is only a rough thought
Raw daily notes are source material. Do not create Wiki pages unless the user explicitly asks.
Final Response
Keep the response short:
已记录到 E:\LLM_wiki\LLM_wiki\01.raw\02.DailyNotes\YYYY-MM-DD_YYYY-MM-DD.md 的 Question-疑问。
If several categories were updated, list them briefly.
Sync Rule
????? skill ???????????? VectorPeak/vectorpeak-agent-skills?????????????
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: VectorPeak
- Source: VectorPeak/vectorpeak-agent-skills
- License: Apache-2.0
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.