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/fabioc-aloha/alex-cognitive-architecture --skill project-registration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。
基于 SOC 职业分类
正在显示 SKILL.md
| name | project-registration |
| description | Register projects in the fleet registry and track cross-project patterns |
| tier | standard |
| applyTo | **/*registry*,**/*fleet*,**/*project-health* |
| currency | 2026-04-22T00:00:00.000Z |
Register and update project metadata in the AI-Memory project registry for fleet-wide visibility and cross-project pattern discovery.
Enable heirs to self-register in the centralized project registry, track successful patterns, document friction points, and promote reusable solutions to the fleet.
Heirs are active contributors to the shared knowledge base, not passive consumers:
| Resource | Heir Can | Example |
|---|---|---|
project-registry.json | Update own entry | Add patterns, friction, health |
insights/ | Create new files | GI-{topic}-{date}.md |
feedback/ | Create new files | Report friction to Master |
user-profile.json | Read only | Identity is centrally managed |
Contribution flow:
successfulPatternsAI-Memory/insights/AI-Memory/patterns/~/OneDrive - Correa Family/AI-Memory/project-registry.json
The registry syncs across all machines via OneDrive. Never create local copies.
{
"path": "C:\\Development\\ProjectName",
"name": "ProjectName",
"description": "Brief project description",
"lastAccessed": "2026-04-16T10:00:00.000Z",
"status": "active|maintenance|archived",
"technologies": ["nodejs", "typescript", "alex-brain"],
"brainVersion": "8.0.0",
"health": {
"skillCount": 0,
"instructionCount": 0,
"lastMeditation": null,
"meditationCount": 0
},
"successfulPatterns"
Heirs operate WITHOUT Master Alex. Each heir maintains its own registry entry using the update-registry.cjs muscle that syncs to the shared AI-Memory.
# First-time registration
node .github/muscles/update-registry.cjs --register
# Update health metrics (run after brain changes)
node .github/muscles/update-registry.cjs
# Record meditation completion
node .github/muscles/update-registry.cjs --meditated
# Add successful pattern
node .github/muscles/update-registry.cjs --pattern "webview-state"
# Document friction point
node .github/muscles/update-registry.cjs --friction "msal-token-refresh"
The muscle:
| Trigger | Action | Command |
|---|---|---|
| First session in project | Self-register | --register |
| After meditation | Record completion | --meditated |
| Solved problem elegantly | Document pattern | --pattern "name" |
| Recurring issue | Document friction | --friction "issue" |
| Brain updated | Refresh health | (no args) |
When Master Alex scans the fleet:
# Master Alex scanner muscle:
node scripts/scan-projects.cjs
For manual registration:
When you solve a problem elegantly and it could help other projects:
{
"successfulPatterns": [
"webview-state-management",
"atomic-file-writes",
"tmdl-parsing"
]
}
Pattern names should be:
When encountering recurring issues:
{
"frictionPoints": [
"msal-token-refresh-race-condition",
"webpack-5-esm-compatibility"
]
}
Before solving a problem from scratch, check if another project already solved it:
// Query pattern
const registry = JSON.parse(fs.readFileSync(registryPath));
const solutions = registry.projects.filter(p =>
p.successfulPatterns.includes('your-problem-domain')
);
When a pattern proves valuable and should become a Master Alex capability:
promotedPatterns arrayAfter meditation or brain-qa runs:
{
"health": {
"skillCount": 176,
"instructionCount": 136,
"lastMeditation": "2026-04-16T10:00:00Z",
"meditationCount": 15
}
}
const typescriptProjects = registry.projects.filter(p =>
p.technologies.includes('typescript') &&
p.status === 'active'
);
// Find who solved SSE streaming
const sseExperts = registry.projects.filter(p =>
p.successfulPatterns.some(pat => pat.includes('sse'))
);
const needsMeditation = registry.projects.filter(p =>
p.status === 'active' &&
!p.health.lastMeditation
);
| Status | Criteria | Registry Update Frequency |
|---|---|---|
active | Modified in last 7 days | On every significant change |
maintenance | Modified in last 30 days | Weekly or on changes |
archived | Not modified in 30+ days | Rarely/never |
When an heir discovers a valuable pattern, contribute it to the shared knowledge base.
<!-- ~/AI-Memory/insights/GI-{topic}-{date}.md -->
---
project: project-name
date: 2026-04-17
category: backend|frontend|architecture|testing|devops
tags: [relevant, searchable, terms]
status: draft|validated
---
# Insight Title
## Context
What problem were you solving?
## Discovery
What pattern/solution did you find?
## Evidence
- Used in: [list files or scenarios]
- Reuse count: 3+
## Recommendation
How should other projects apply this?
After contributing an insight, update your registry entry:
// Mark that you contributed this pattern
myEntry.successfulPatterns.push('pattern-name');
Master Alex reviews AI-Memory/insights/ during meditation:
AI-Memory/patterns/GK-*.mdThe project registry powers:
master-wiki/architecture/MEMORY-SYSTEMS.md — Full memory architecture (master-only)