By inno-bit-lab
IBL ABP Framework skill bundle for backend, modular architecture, MongoDB, multitenancy, React UI, and testing workflows.
Core ABP Framework knowledge for .NET projects: module system, dependency injection, base classes, audit entities, BusinessException with localized DomainErrorCodes, ConcurrencyStamp, localization, soft delete, multi-tenancy basics, and common anti-patterns. Use whenever the user works on ABP modules, ApplicationService, AggregateRoot, IRepository, IClock, CurrentUser, GuidGenerator, ITransientDependency, auto API controllers, or when another abp-* skill needs shared project detection through abp_context.py. Trigger on any ABP-specific work.
End-to-end ABP feature development with an interview-driven workflow. Use to add or port entities, create AppServices and DTOs, scaffold CRUD endpoints, design aggregates or DomainServices, configure permissions, mapping, filters, bulk delete, Excel export, lookup endpoints, concurrency stamps, soft delete, lifecycle methods, explicit test coverage decisions, and final validation. Trigger on "add an entity", "port this table", "new AppService", "scaffold CRUD", "add permissions", "aggregate root", "DTO", "Mapperly", "filtri di lista", "export Excel", or any ABP feature work.
Use when an ABP project is or should become modular: deciding entity/module ownership, proposing Shared or bounded-context modules, moving backend/React/localization/permissions between modules, or diagnosing missing menu/styles/permissions after modularization.
ABP Framework + MongoDB integration: configure AbpMongoDbContext, register entity collections, add custom MongoDbRepository implementations, typed filters, indexes, embedded/reference documents, enum string setup, and explicit UpdateAsync behavior because MongoDB has no change tracking. Use when adding a Mongo-backed ABP entity or working with MongoDbContext, IMongoCollection, MongoDbRepository, IMongoDbContextProvider, AddMongoDbContext, GetQueryableAsync, GetCollectionAsync, indexes, compound indexes, tenant-aware uniqueness, or collection registration.
ABP Framework multi-tenancy - making entities tenant-aware via IMultiTenant, accessing CurrentTenant, switching tenant context with CurrentTenant.Change, bypassing the multi-tenant data filter, tenant resolution strategies, and host vs tenant behavior. Use to add multi-tenancy to an entity, query across tenants, switch tenant or host context, configure tenant resolution, define permissions per tenancy side, or troubleshoot tenant-filter visibility issues. Triggers on IMultiTenant, TenantId, CurrentTenant, multi-tenant, switch tenant, data filter, DataFilter.Disable, or tenant resolver.
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimBased on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Repository sorgente per plugin e skill aziendali riusabili da agenti AI.
Questo repo è pensato per vivere su GitHub ed essere il punto canonico da cui Codex, Claude Code, Antigravity e OpenCode ricevono plugin, skill e miglioramenti.
Il contenuto delle skill vive una sola volta:
plugins/<plugin>/skills/<skill>/
Le piattaforme non devono mantenere testi divergenti. I manifest e gli script sono adattatori per host diversi; il sorgente autorevole resta questo repo.
plugins/ibl-abp - skill ABP Framework per backend .NET, MongoDB, multitenancy, React UI e test.plugins/ibl-skill-improvement - skill per catturare, revisionare, applicare improvement e aggiornare il checkout dei plugin aziendali.plugins/ibl-agent-workflow - skill di workflow cross-agent: scrivere un documento di handoff di sessione (redatto, salvato nella temp dir del SO) perché un agente fresco possa continuare il lavoro..
├── .agents/plugins/marketplace.json # marketplace Codex locale al repo
├── .claude-plugin/marketplace.json # marketplace Claude Code locale al repo
├── improvements/ # inbox degli improvement prodotti dagli agenti
├── plugins/
│ ├── ibl-abp/
│ │ ├── .codex-plugin/plugin.json # manifest Codex
│ │ ├── .claude-plugin/plugin.json # manifest Claude Code
│ │ ├── plugin.json # manifest Antigravity
│ │ └── skills/ # sorgente unico delle skill
│ └── ibl-skill-improvement/
│ ├── .codex-plugin/plugin.json
│ ├── .claude-plugin/plugin.json
│ ├── plugin.json
│ └── skills/
└── tools/
Ogni plugin deve esporre, quando applicabile:
plugins/<plugin>/
├── .codex-plugin/plugin.json
├── .claude-plugin/plugin.json
├── plugin.json
└── skills/
OpenCode consuma direttamente le cartelle skills/<skill> tramite link o copia
controllata, senza duplicare il contenuto nel repository.
Quando una skill sbaglia, l'agent deve creare una proposta versionata in:
improvements/inbox/<id>/
La proposta contiene:
problem.mdimprovement.mdmodified-resources.mdvalidation.mdcandidate/ opzionale con la versione migliorata dei fileIl plugin ibl-skill-improvement lavora questa inbox, revisiona la proposta e
applica le correzioni solo al sorgente canonico sotto plugins/.
Quando la skill è installata in Codex, Claude Code, Antigravity o OpenCode,
l'area di installazione è solo una cache. Gli improvement devono comunque essere
scritti nel checkout Git canonico. Il marketplace non imposta automaticamente
IBL_AGENT_PLUGINS_HOME: quella variabile è solo un override opzionale per path
non standard.
Il helper skills/skill-improvement/scripts/improvement_inbox.py cerca il
checkout tramite directory corrente, path standard sotto
%USERPROFILE%\agent-marketplaces\, installazioni linkate e, se configurato,
IBL_AGENT_PLUGINS_HOME. Se non trova il checkout, mostra il comando gh repo clone da eseguire una volta.
Su GitHub il flusso previsto è:
skill fallisce
-> artifact in improvements/inbox
-> branch proposal-only
-> PR proposal-only
-> process/apply PR
-> review e validazione della modifica applicata
-> merge
-> update degli agent dal repo
In Capture Mode l'agent che propone l'improvement deve pubblicare la richiesta, non applicarla automaticamente:
python plugins/ibl-skill-improvement/skills/skill-improvement/scripts/improvement_inbox.py publish improvements/inbox/<artifact-id> --create-pr
Questa operazione committa e pusha solo improvements/inbox/<artifact-id>/.
Le modifiche reali sotto plugins/ sono responsabilità del passaggio Process
Mode, che può essere eseguito da un altro agent o maintainer.
Gli artifact applicati vengono spostati in improvements/applied/. Gli artifact
non validi o superati vengono spostati in improvements/rejected/.
Il canale di update del team è GitHub. Se il repo è già clonato, il fallback manuale resta:
cd "$env:USERPROFILE\agent-marketplaces\ibl-agent-plugins"
git pull --ff-only
Con il plugin ibl-skill-improvement installato, l'utente può invece chiedere
all'agent di aggiornare i plugin IBL senza ricordare la cartella. La skill
agent-plugin-update usa questo helper:
python plugins/ibl-skill-improvement/skills/agent-plugin-update/scripts/update_agent_plugins.py --validate
L'helper cerca il checkout tramite directory corrente, path standard sotto
%USERPROFILE%\agent-marketplaces\, installazioni linkate di Antigravity,
Claude Code o OpenCode e, se configurato, IBL_AGENT_PLUGINS_HOME.
Per Codex, se il marketplace è stato aggiunto direttamente da GitHub:
codex plugin marketplace add inno-bit-lab/ibl-agent-plugins --ref main
la snapshot Codex si aggiorna con:
npx claudepluginhub inno-bit-lab/ibl-agent-plugins --plugin ibl-abpIBL workflow for capturing, reviewing, applying, and updating reusable agent skills.
IBL cross-agent workflow skills: write a session-handoff document so a fresh agent can continue the work.
A growing collection of Claude-compatible academic workflow bundles. Covers scientific figures, manuscript writing and polishing, reviewer assessment, citation retrieval, data availability, paper reading, literature search, response letters, paper-to-PPTX conversion, and evidence-grounded Chinese invention patent drafting. Rules are organized as reusable skill folders with explicit workflows and quality checks.
Comprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Tools to maintain and improve CLAUDE.md files - audit quality, capture session learnings, and keep project memory current.
Persistent file-based planning for AI coding agents. Crash-proof markdown plans (task_plan.md, findings.md, progress.md) that survive context loss and /clear, with an opt-in completion gate and multi-agent shared state. Manus-style. Works with Claude Code, Codex CLI, Cursor, Kiro, OpenCode and 60+ agents via the SKILL.md standard. Includes Arabic, German, Spanish, and Chinese (Simplified and Traditional).
Browser automation and end-to-end testing MCP server by Microsoft. Enables Claude to interact with web pages, take screenshots, fill forms, click elements, and perform automated browser testing workflows.
Comprehensive feature development workflow with specialized agents for codebase exploration, architecture design, and quality review