diary
Unified Diary System: A context-preserving automated logger for multi-project development.
Best use case
diary is best used when you need a repeatable AI agent workflow instead of a one-off prompt.
Unified Diary System: A context-preserving automated logger for multi-project development.
Teams using diary should expect a more consistent output, faster repeated execution, less prompt rewriting.
When to use this skill
- You want a reusable workflow that can be run more than once with consistent structure.
When not to use this skill
- You only need a quick one-off answer and do not need a reusable workflow.
- You cannot install or maintain the underlying files, dependencies, or repository context.
Installation
Claude Code / Cursor / Codex
Manual Installation
- Download SKILL.md from GitHub
- Place it in
.claude/skills/diary/SKILL.mdinside your project - Restart your AI agent ā it will auto-discover the skill
How diary Compares
| Feature / Agent | diary | Standard Approach |
|---|---|---|
| Platform Support | Not specified | Limited / Varies |
| Context Awareness | High | Baseline |
| Installation Complexity | Unknown | N/A |
Frequently Asked Questions
What does this skill do?
Unified Diary System: A context-preserving automated logger for multi-project development.
Where can I find the source code?
You can find the source code on GitHub using the link provided at the top of the page.
Related Guides
AI Agents for Coding
Browse AI agent skills for coding, debugging, testing, refactoring, code review, and developer workflows across Claude, Cursor, and Codex.
Cursor vs Codex for AI Workflows
Compare Cursor and Codex for AI coding workflows, repository assistance, debugging, refactoring, and reusable developer skills.
AI Agents for Marketing
Discover AI agents for marketing workflows, from SEO and content production to campaign research, outreach, and analytics.
SKILL.md Source
# š Unified Diary System
## When to Use This Skill
Use this skill when you want to summarize progress, write a daily dev log, or perform a daily review while keeping project contexts isolated and synced to Notion/Obsidian.
> šØ **Agent One-Shot Integrity Constraint (Highest Priority)**: Steps 1-4 are an **indivisible atomic workflow**. The AI **MUST use Continuous Tool Calling** to complete all actions in one breath.
> - **ABSOLUTELY FORBIDDEN** to output conversational text and wait for the user after completing Step 1, 2, or 3.
> - Example: Immediately after saving the file in Step 1, you must **automatically and instantly** call the terminal tool to execute the script in Step 2. Stopping to chat or asking "Should I continue?" is considered a critical error. Only inform the user explicitly if you are stuck due to a technical error.
> ā ļø **Ironclad Rule 1**: The local project diary in Step 1 **must NEVER be polluted with global data**. It only records the current project.
> ā ļø **Ironclad Rule 2**: When using tools like `run_command` to execute predefined python scripts (Step 2, Step 4), you must set them to `SafeToAutoRun: true` to avoid getting stuck waiting for user authorization.
> ā ļø **Ironclad Rule 3**: **Never guess the project name**. The AI must confirm the current directory name as the project identifier using a terminal command BEFORE taking any action in Step 1.
---
## Step 1: Local Project Archiving (AI Generation)
- **Action 0 (Identify)**: Call the terminal `pwd` (Linux/Mac) or `(Get-Item .).Name` (Windows) to confirm the current folder name.
- **Action 1 (Write)**: Summarize the achievements from the current conversation (Git Commits, file changes, task progress), and write them into the **current project folder** at `diary/YYYY/MM/YYYY-MM-DD-ProjectName.md`.
- **Isolation and Naming Rules (Ironclad Rules)**:
- š **Mandatory Filename Suffix**: The local diary **MUST** include the project name detected just now. It is **absolutely forbidden** to use a global-level filename (like `2026-02-23.md`) locally.
- ā
**Pure Content**: Only record content exclusive to the current project. Do not mix in other projects.
- š **Append Mode**: If the project diary already exists, update it using "append", never overwrite the original content.
- š **Auto-Creation**: Create subfolders `diary/YYYY/MM/` based on the year and month.
- ā” **Force Continue**: Once writing is complete, **do not interrupt the conversation; immediately call the terminal tool and proceed to Step 2.**
## Step 1.5: Refresh Project Context (Automation Script)
- **Prerequisite**: You have confirmed the current project directory path (from Action 0's `pwd` result).
- **Action**: Call the terminal to execute the following command to automatically scan the project state and generate/update `AGENT_CONTEXT.md`:
```powershell
python {diary_system_path}/scripts/prepare_context.py "<Project_Root_Path>"
```
- **SafeToAutoRun**: true (Safe operation; purely reading and writing local files).
- **Result**: `AGENT_CONTEXT.md` in the project directory is refreshed to the latest state.
- **After Completion**: Force continue to Step 2; do not wait for user confirmation.
## Step 2: Extract Global & Project Material (Script Execution)
- **Action**: Call the extraction script, **passing in the absolute path of the project diary just written in Step 1**. The script will precisely print "Today's Global Progress" and "Current Project Progress".
- **Execution Command**:
```powershell
python {diary_system_path}/scripts/fetch_diaries.py "<Absolute_Path_to_Step1_Project_Diary>"
```
- **Result**: The terminal will print two sets of material side-by-side. The AI must read the terminal output directly and prepare for mental fusion.
## Step 3: AI Smart Fusion & Global Archiving (AI Execution) š§
- **Action**: Based on the two materials printed by the terminal in Step 2, complete a **seamless fusion** mentally, then write it to the global diary: `{diary_system_path}/diary/YYYY/MM/YYYY-MM-DD.md`.
- **Context Firewall (Core Mechanism)**:
1. **No Tag Drift**: When reading "Global Progress Material", there may be progress from other projects. **It is strictly forbidden to categorize today's conversation achievements under existing project headings belonging to other projects.**
2. **Priority Definition**: The content marked as `š [Current Project Latest Progress]` in Step 2 is the protagonist of today's diary.
- **Rewrite Rules**:
1. **Safety First**: If the global diary "already exists," preserve the original content and append/fuse the new project progress. **Do not overwrite.**
2. **Precise Zoning**: Ensure there is a dedicated `### š ProjectName` zone for this project. Do not mix content into other project zones.
3. **Lessons Learned**: Merge and deduplicate; attach action items to every entry.
4. **Cleanup**: After writing or fusing globally, you **must** force-delete any temporary files created to avoid encoding issues (e.g., `temp_diary.txt`, `fetched_diary.txt`) to keep the workspace clean.
## Step 4: Cloud Sync & Experience Extraction (Script + Human) š
- **Action 1 (Sync)**: Call the master script to push the global diary to Notion and Obsidian.
- **Execution Command**:
```powershell
python {diary_system_path}/scripts/master_diary_sync.py --sync-only
```
- **Action 2 (Extraction & Forced Pause)**:
1. The AI extracts "Improvements & Learning" from the global diary.
2. Confirm if it contains entirely new key points lacking in the past (š New Rules), or better approaches (š Evolved Rules).
3. List the results and **WAIT FOR USER CONFIRMATION** (user says "execute" or "agree").
4. After user confirmation, update the `.md` file in `{Knowledge_Base_Path}/` and execute `qmd embed` (if applicable).
---
**šÆ Task Acceptance Criteria**:
1. ā
Project local diary generated (no pollution).
2. ā
`fetch_diaries.py` called with absolute path and successfully printed materials.
3. ā
AI executed high-quality rewrite and precisely wrote to global diary (appended successfully if file existed).
4. ā
`--sync-only` successfully pushed to Notion + Obsidian.
5. ā
Experience extraction presented to the user and authorized.
---
## š Templates and Writing Guidelines
Strictly apply the following Markdown templates to ensure clarity during Step 1 (Local) and Step 3 (Global Fusion).
### š” Writing Guidelines (For AI)
1. **Dynamic Replacement**: The `{Project Name}` in the template MUST strictly use the folder name grabbed by `pwd` in Step 1.
2. **Concise Deduplication**: When writing the global diary in Step 3, the AI must condense the "š ļø Execution Details" from the local diary. The global diary focuses only on "General Direction and Output Results."
3. **Mandatory Checkboxes**: All "Next Steps" and "Action Items" must use the Markdown `* [ ]` format so they can be checked off in Obsidian/Notion later.
### š Template 1: Project Local Diary (Step 1 Exclusive)
```markdown
# Project DevLog: {Project Name}
* **š
Date**: YYYY-MM-DD
* **š·ļø Tags**: `#Project` `#DevLog`
---
> šÆ **Progress Summary**
> (Briefly state the core task completed, e.g., "Finished Google Colab environment testing for auto-video-editor")
### š ļø Execution Details & Changes
* **Git Commits**: (List if any)
* **Core File Modifications**:
* š `path/filename`: Explanation of changes.
* **Technical Implementation**:
* (Record key logic or architecture structural changes)
### šØ Troubleshooting
> š **Problem Encountered**: (e.g., API error, package conflict)
> š” **Solution**: (Final fix, leave key commands)
### āļø Next Steps
- [ ] (Specific task 1)
- [ ] (Specific task 2)
```
---
### š Template 2: Global Diary (Step 3 Exclusive)
```markdown
# š YYYY-MM-DD Global Progress Overview
> š **Daily Highlight**
> (1-2 sentences summarizing all project progress for the day, synthesized by AI)
---
## š Project Tracking
(ā ļø AI Rule: If file exists, find the corresponding project title and append; NEVER overwrite, keep it clean.)
### šµ {Project A, e.g., auto-video-editor}
* **Today's Progress**: (Condense Step 2 local materials into key points)
* **Action Items**: (Extract next steps)
### š¢ {Project B, e.g., GSS}
* **Today's Progress**: (Condense key points)
* **Action Items**: (Extract next steps)
---
## š§ Improvements & Learnings
(ā ļø Dedicated to Experience Extraction)
š **New Rules / Discoveries**
(e.g., Found hidden API limit, or a more efficient python syntax)
š **Optimizations & Reflections**
(Improvements from past methods)
---
## ā
Global Action Items
- [ ] (Tasks unrelated to specific projects)
- [ ] (System environment maintenance, etc.)
```Related Skills
using-git-worktrees
Git worktrees create isolated workspaces sharing the same repository, allowing work on multiple branches simultaneously without switching.
slack-gif-creator
A toolkit providing utilities and knowledge for creating animated GIFs optimized for Slack.
skill-creator
To create new CLI skills following Anthropic's official best practices with zero manual configuration. This skill automates brainstorming, template application, validation, and installation processes while maintaining progressive disclosure patterns and writing style standards.
project-skill-audit
Audit a project and recommend the highest-value skills to add or update.
notebooklm
Interact with Google NotebookLM to query documentation with Gemini's source-grounded answers. Each question opens a fresh browser session, retrieves the answer exclusively from your uploaded documents, and closes.
molykit
CRITICAL: Use for MolyKit AI chat toolkit. Triggers on: BotClient, OpenAI, SSE streaming, AI chat, molykit, PlatformSend, spawn(), ThreadToken, cross-platform async, Chat widget, Messages, PromptInput, Avatar, LLM
claude-settings-audit
Analyze a repository to generate recommended Claude Code settings.json permissions. Use when setting up a new project, auditing existing settings, or determining which read-only bash commands to allow. Detects tech stack, build tools, and monorepo structure.
claude-in-chrome-troubleshooting
Diagnose and fix Claude in Chrome MCP extension connectivity issues. Use when mcp__claude-in-chrome__* tools fail, return "Browser extension is not connected", or behave erratically.
claude-code-expert
Especialista profundo em Claude Code - CLI da Anthropic. Maximiza produtividade com atalhos, hooks, MCPs, configuracoes avancadas, workflows, CLAUDE.md, memoria, sub-agentes, permissoes e integracao com ecossistemas.
blockrun
BlockRun works with Claude Code and Google Antigravity.
auri-core
Auri: assistente de voz inteligente (Alexa + Claude claude-opus-4-20250805). Visao do produto, persona Vitoria Neural, stack AWS, modelo Free/Pro/Business/Enterprise, roadmap 4 fases, GTM, north star WAC e analise competitiva.
audit-skills
Expert security auditor for AI Skills and Bundles. Performs non-intrusive static analysis to identify malicious patterns, data leaks, system stability risks, and obfuscated payloads across Windows, macOS, Linux/Unix, and Mobile (Android/iOS).