Ai Agent Setup Script

D 54 completed
Framework
monorepo / typescript · small
194
Files
12,738
LOC
2
Frameworks
8
Languages

Pipeline State

completed
Run ID
#93319
Phase
done
Progress
0%
Started
2026-03-21 00:53:24
Finished
2026-03-21 00:53:24
LLM tokens
0

Pipeline Metadata

Stage
Cataloged
Decision
proceed
Novelty
75.00
Framework unique
Isolation
Last stage change
2026-05-10 03:35:10
Deduplication group #53801
Member of a group with 1 similar repo(s) — this repo is canonical view group →
Top concepts (2)
Project DescriptionFull Stack
Methodology: Repobility · https://repobility.com/research/state-of-ai-code-2026/

AI Prompt

Build me a personal, secure AI agent framework called Murph. I need it to be able to communicate across multiple channels like iMessage, Telegram, and a local Next.js web dashboard. The core functionality should include a 3-tier memory system using PostgreSQL and pgvector for semantic search, and a knowledge base that can ingest data from Obsidian vaults, PDFs, and web pages. Crucially, every action must pass through configurable approval gates, and all activity needs to be audited and logged to PostgreSQL. Also, implement secure secret management using AES-256-GCM stored in the macOS Keychain.
typescript next.js ai-agent monorepo postgresql pgvector imessage telegram security automation
Generated by gemma4:latest

Catalog Information

A framework for building a persistent, secure AI assistant that can communicate via iMessage, Telegram, or a web dashboard.

Description

This framework enables developers to create a personal AI assistant that runs on a Mac Mini and interacts through multiple channels such as iMessage, Telegram, and a local web dashboard. It integrates a three‑tier memory system—short‑term buffer, PostgreSQL long‑term storage, and pgvector semantic search—to provide context‑aware responses. The assistant can ingest and search a wide range of knowledge sources, including Obsidian vaults, PDFs, web pages, and voice notes, while enforcing configurable approval gates and detailed audit logging. Security is handled with AES‑256‑GCM encryption and macOS Keychain for secret storage, and the system supports natural‑language scheduling, dynamic cloud deployment, and various third‑party integrations. The design focuses on extensibility, allowing developers to add new channels, integrations, or custom actions with minimal effort.

الوصف

يُقدّم إطار عمل لبناء مساعد ذكاء اصطناعي شخصي يعمل على جهاز Mac Mini ويتفاعل عبر قنوات متعددة مثل iMessage وTelegram ولوحة تحكم ويب محلية. يدمج نظام ذاكرة ثلاثي المستويات—الذاكرة القصيرة الأجل، تخزين PostgreSQL طويل الأجل، والبحث الدلالي عبر pgvector—لتوفير ردود ذات سياق دقيق. يمكن للمساعد استيعاب والبحث في مصادر معرفة واسعة، بما في ذلك مستودعات Obsidian، ملفات PDF، صفحات ويب، وملاحظات صوتية، مع تطبيق بوابات موافقة قابلة للتكوين وتسجيل تدقيق مفصل. يتم تأمين النظام باستخدام تشفير AES‑256‑GCM وتخزين الأسرار في Keychain الخاص بـ macOS، كما يدعم جدولة المهام باستخدام اللغة الطبيعية، ونشر التطبيقات السحابي الديناميكي، وتكاملات طرف ثالث متعددة. يركز التصميم على القابلية للتوسع، مما يتيح للمطورين إضافة قنوات جديدة، تكاملات، أو إجراءات مخصصة بسهولة.

Novelty

8/10

Tags

ai-agent multi-channel-messaging persistent-memory knowledge-base-ingestion approval-workflow audit-logging encrypted-secrets scheduler

Claude Models

claude-opus-4.6

Quality Score

D
54.3/100
Structure
41
Code Quality
88
Documentation
36
Testing
0
Practices
74
Security
75
Dependencies
60

Weaknesses

  • No LICENSE file \u2014 legal ambiguity for contributors
  • No tests found \u2014 high risk of regressions
  • No CI/CD configuration \u2014 manual testing and deployment
  • 1 files with critical complexity need refactoring
  • 657 duplicate lines detected \u2014 consider DRY refactoring

Recommendations

  • Add a test suite \u2014 start with critical path integration tests
  • Set up CI/CD (GitHub Actions recommended) to automate testing and deployment
  • Add a linter configuration to enforce code style consistency
  • Add a LICENSE file (MIT recommended for open source)
  • Address 23 TODO/FIXME items \u2014 consider tracking them as issues

Security & Health

1
Vulnerabilities
0
Critical CVEs
15.8h
Tech Debt (C)
A
OWASP (100%)
Repobility analyzer · published findings · https://repobility.com
PASS
Quality Gate
A
Risk (3)
Unknown
License
4.5%
Duplication
Full Security Report AI Fix Prompts SARIF SBOM

Languages

typescript
65.4%
yaml
23.3%
json
5.1%
shell
3.6%
markdown
1.7%
sql
0.8%
javascript
0.0%
css
0.0%

Frameworks

Next.js Turbopack

Concepts (2)

Analysis by Repobility (https://repobility.com) · MCP-ready
CategoryNameDescriptionConfidence
Repobility · MCP-ready · https://repobility.com
auto_descriptionProject DescriptionA personal AI agent framework built on top of Claude Code CLI. Murph runs on a Mac Mini and provides a persistent, secure, extensible assistant reachable via iMessage, Telegram, or a local web dashboard.80%
auto_categoryFull Stackfull-stack70%

Threat Findings

Repobility's GitHub App fixes findings like these · https://github.com/apps/repobility-bot
1
Total Threats
0
Critical
0
High
View Threat Intel

Quality Timeline

1 quality score recorded.

View File Metrics

Embed Badge

Add to your README:

![Quality](https://repos.aljefra.com/badge/92545.svg)
Quality BadgeSecurity Badge
Export Quality CSVDownload SBOMExport Findings CSV