Skip to content

AGENTS.md

Latest commit

 

History

History
115 lines (96 loc) · 3.95 KB

AGENTS.md

File metadata and controls

115 lines (96 loc) · 3.95 KB

PROJECT KNOWLEDGE BASE

Project: oh-my-onpremcode Author: sionic-ai Description: On-premise OpenCode plugin with local inference support for air-gapped environments

OVERVIEW

Fork of oh-my-opencode designed for enterprise on-premise deployments. Supports vLLM (Kubernetes) and Ollama for local inference. All internet-dependent features auto-disable in on-prem mode.

STRUCTURE

oh-my-onpremcode/
├── src/
│   ├── agents/        # AI agents (7): Sisyphus, oracle, librarian, explore, frontend, document-writer, multimodal-looker
│   ├── hooks/         # 21 lifecycle hooks - see src/hooks/AGENTS.md
│   ├── tools/         # LSP, AST-Grep, Grep, Glob, etc. - see src/tools/AGENTS.md
│   ├── mcp/           # MCP servers (disabled in onprem mode)
│   ├── inference/     # Local inference clients (vLLM, Ollama)
│   ├── features/      # Claude Code compatibility - see src/features/AGENTS.md
│   ├── config/        # Zod schema, TypeScript types
│   ├── auth/          # Google OAuth (disabled in onprem mode)
│   ├── shared/        # Utilities: deep-merge, pattern-matcher, logger, project-info
│   └── index.ts       # Main plugin entry
├── script/            # build-schema.ts, publish.ts, generate-changelog.ts
├── assets/            # JSON schema
└── dist/              # Build output (ESM + .d.ts)

KEY FILES FOR ON-PREM

File Purpose
src/inference/vllm.ts vLLM client (OpenAI-compatible API)
src/inference/ollama.ts Ollama client
src/inference/index.ts InferenceManager with auto-failover
src/config/schema.ts OnPremConfigSchema, VllmConfigSchema, etc.
src/shared/project-info.ts Centralized project constants

WHERE TO LOOK

Task Location Notes
Add agent src/agents/ Create .ts, add to builtinAgents in index.ts
Add hook src/hooks/ Create dir with createXXXHook()
Add tool src/tools/ Dir with index/types/constants/tools.ts
Modify inference src/inference/ vllm.ts, ollama.ts, index.ts
Config schema src/config/schema.ts Run bun run build:schema after changes
Project constants src/shared/project-info.ts Package name, org, URLs

CONVENTIONS

  • Package manager: Bun only (bun run, bun build, bunx)
  • Types: bun-types (not @types/node)
  • Build: Dual output - bun build (ESM) + tsc --emitDeclarationOnly
  • Exports: Barrel pattern - export * from "./module" in index.ts
  • Directory naming: kebab-case (ast-grep/, claude-code-hooks/)
  • Config file: oh-my-onpremcode.json

ON-PREM CONFIGURATION

{
  "onprem": {
    "mode": "onprem",
    "vllm": {
      "endpoints": [{ "url": "http://vllm:8000/v1", "model": "llama-3-70b" }],
      "k8s_namespace": "default",
      "k8s_service_name": "vllm-service",
      "auto_discover": true
    },
    "ollama": {
      "enabled": true,
      "host": "localhost",
      "port": 11434
    },
    "model_mapping": {
      "anthropic/claude-opus-4-5": "meta-llama/Llama-3.1-70B-Instruct"
    },
    "branding": {
      "company_name": "Sionic AI",
      "welcome_message": "Welcome to Sionic AI"
    }
  }
}

DISABLED IN ON-PREM MODE

Feature Reason
context7 MCP Requires internet
websearch_exa MCP Requires internet
grep_app MCP Requires internet
Google OAuth Requires internet
Auto-update Requires npm registry

COMMANDS

bun install            # Install dependencies
bun run build          # ESM + declarations + schema
bun run build:schema   # Schema only
bun test               # Run tests

NOTES

  • Config locations:
    • User: ~/.config/opencode/oh-my-onpremcode.json
    • Project: .opencode/oh-my-onpremcode.json
  • OpenCode: Requires >= 1.0.150
  • vLLM: OpenAI-compatible API on /v1/chat/completions
  • Ollama: Native API on /api/chat