メニュー
Create new files in knowledge base with directory governance. Use when user wants to save, create, or add content. Reads RULE.md to understand directory purpose and follows specified rules. Recommends target directory based on content analysis.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
SOC 職業分類に基づく
Self-iterating autonomous loop where the main agent acts as a pure orchestrator, spawning isolated Implementor and Critic subagents in rounds until the artifact converges, an iteration cap is reached, or human escalation is needed. Domain-agnostic — works for code, documents, plans, designs, copy, specs, proposals, or anything with an iterable artifact and explicit acceptance criteria. Also addressable by the short alias "bvl".
Initialize the Engram memory system — creates configuration, memory files, and CLAUDE.md integration. Trigger when the user sets up persistent memory for a project, or when `.claude/memory-settings.json` is missing and a memory feature is requested.
Search and recall memories from the Engram files. Trigger when a topic may relate to past discussions, when the user references previous context, or when the user asks "do you remember…".
Store a memory into the appropriate Engram file. Trigger when a new preference, decision, milestone, or natural conversation summary point is identified, or when the user explicitly asks to remember something.
Display Engram memory system status — configuration, file sizes, entry counts, and health check. Trigger when the user asks about memory state, what's been remembered, or for diagnostics.
Manage the dynamic skill library. Initialize skill-library/ and SKILLUSE.md, add new skills, remove skills, edit existing skills, and list available skills. Use when the user wants to manage, create, or modify skills in the skill library.
| name | add-content |
| description | Create new files in knowledge base with directory governance. Use when user wants to save, create, or add content. Reads RULE.md to understand directory purpose and follows specified rules. Recommends target directory based on content analysis. |
Generic file creation Skill that works with any directory structure by reading RULE.md to understand purpose and rules.
Determine content type:
Extract key information:
Scan knowledge base:
ls or GlobExample scan result:
ReadLater/ - Purpose: "Archive web content for later reading"
Research/ - Purpose: "Technical and academic research notes"
Work/ - Purpose: "Professional work-related content"
Personal/ - Purpose: "Personal life documentation"
Miscellaneous/ - Purpose: "General notes and articles"
Match content to directories:
Classification logic:
Present recommendation:
I analyzed your content about [topic].
Recommended location: [Directory]/
Reason: [Why this directory suits the content]
Is this correct?
- Yes (proceed with this directory)
- No, use different directory (please specify path)
- Show me other options
If user requests alternatives:
If user specifies custom path:
Locate and read RULE.md:
Parse RULE.md for:
YYYY-MM-DD-title.md)Follow RULE.md instructions exactly:
Example 1 - Simple Note:
RULE.md says: "Files named YYYY-MM-DD-title.md with markdown format"
→ Generate filename: 2025-10-28-transformer-architecture.md
→ Create file with content
→ Save in target directory
Example 2 - Web Archival:
RULE.md says:
"When user provides URL:
1. Fetch content using WebFetch
2. Convert to markdown
3. Create directory: Articles/YYYY/MM/YYYY-MM-DD_title-slug/
4. Save article.md and metadata.yaml"
→ Execute WebFetch on URL
→ Extract title, content, date
→ Convert to clean markdown
→ Create directory structure
→ Save article.md with content
→ Create metadata.yaml with:
- url: [original URL]
- title: [article title]
- date: [fetch date]
- summary: [brief summary]
→ Optionally save images if referenced
Example 3 - Structured Content (WorkLog):
RULE.md says:
"Daily logs in YYYY/QX/MM-Month/ structure.
MUST verify datetime first.
Format: [specific template shown in RULE.md]"
→ Execute `date` command to verify current date
→ Parse date: 2025-10-28 → 2025/Q4/10-October/
→ Check if directory structure exists, create if needed
→ Read RULE.md template section
→ Create/update 2025-10-28.md with template format
→ Fill in sections per RULE.md specifications
Key principles:
Update README.md:
Update parent README.md if needed:
Verify updates:
Confirm operation complete:
✅ Content created successfully
Location: [full path to file]
Format: [file format used]
Governance: [which RULE.md applied]
README.md updated:
- [target directory]/README.md
- [parent directory]/README.md (if applicable)
Summary:
[Brief summary of what was created and where]
When content is URL and RULE.md specifies archival:
Validate URL:
Check for Duplicates:
Fetch Content:
Process Content:
Save Per RULE.md:
If RULE.md requires YAML frontmatter:
---
title: [Title]
date: [Date]
tags: [tag1, tag2]
category: [Category]
---
[Content here]
Generate frontmatter with:
If RULE.md specifies creating multiple files:
Parse RULE.md for custom instructions:
Pattern detection:
Execute custom logic:
This Skill automatically invokes the governance protocol:
Before operation:
During operation:
After operation:
User: "Save this note: Today I learned about transformer architecture in deep learning"
Skill workflow:
YYYY-MM-DD-topic.md2025-10-28-transformer-architecture.mdUser: "readlater https://example.com/interesting-article"
Skill workflow:
ReadLater/Articles/2025/10/2025-10-28_interesting-article-about-ai/User: "Update my worklog"
Skill workflow:
date → 2025-10-28