rag-mcp-lsp-ops
Handles operational setup and health checks for RAG, MCP, and LSP tooling.
Install
mkdir -p .claude/skills/rag-mcp-lsp-ops && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/16598" && unzip -o skill.zip -d .claude/skills/rag-mcp-lsp-ops && rm skill.zipInstalls to .claude/skills/rag-mcp-lsp-ops
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Use this skill when tasks involve RAG/MCP/LSP operations in this repository: bringing PM2 + MCPKey capabilities
- →Check Ollama local status
- →Ensure PM2 runtime is active
- →Diagnose MCP tool availability
- →Rebuild RAG index from zero
- →Validate usable retrieval with `rag_search`
- →Validate LSP semantic tooling readiness
How it works
This skill defines the operational workflow for RAG, MCP, and LSP by performing preflight runtime checks, diagnosing MCP, rebuilding RAG, and validating retrieval and index freshness. It ensures semantic navigation and code context retrieval are ready for coding workflows.
Inputs & outputs
When to use rag-mcp-lsp-ops
- →Rebuilding RAG index from zero
- →Diagnosing MCP availability
- →Validating LSP semantic tooling
- →Preparing PM2 runtime environment
About this skill
RAG MCP LSP Ops
Overview
This skill defines the canonical operational workflow for RAG + MCP + LSP in this repository. Use it when the assistant needs reliable code-context retrieval, semantic navigation readiness, or full RAG rebuild. For contract governance, runbook observability and deep proposal generation, combine with:
audit-contracts-v3-opsaudit-runbook-observabilityaudit-proposal-deep-triage
When specifically working with language-server features (definitions, completions, diagnostics), you
can also pair this with the dedicated lsp-ops skill for finer semantic guidance.
When To Use
- User asks to rebuild/reindex RAG.
- MCP tools are unavailable or inconsistent.
- LSP tools need validation before deeper coding work.
- Assistant must prepare runtime (
pm2,server,mcp) before coding tasks that depend on tool context.
Canonical Workflow
- Preflight runtime.
- Check Ollama local:
curl http://host.docker.internal:11434/api/version - Ensure PM2 runtime:
npm run daemon:start - Wait endpoints:
GET /healthandGET /api/mcponhttp://localhost:3008
- Diagnose MCP.
- Run
npm run mcp:diagnose - Confirm required tools (
rag_*,ollama_*,lsp_*) are listed.
- Rebuild RAG from zero.
- Run
npm run rag:rebuild:zero - If needed, force profile full manually:
npm run rag:index -- --profile full - For scope control, use the same flags across rebuild/index/watch:
--docs-mode include|exclude|only--include-glob/--exclude-glob(repeatable)--max-file-bytes
- Recommended phased indexing:
- Code/config first (no markdown):
npm run rag:index -- --profile full --docs-mode exclude - Docs later:
npm run rag:index -- --profile full --docs-mode only
- Code/config first (no markdown):
- Validate usable retrieval.
- Execute one
rag_searchsmoke call over MCP (tools/call) and confirm non-error response.
- Validate index freshness policy.
- Prefer index freshness within 24h for high-signal triage.
- If
index_freshness_msis stale, mark reduced confidence and recommend refresh/rebuild.
- Continue coding flow.
- For normal coding queries use
profile=core. - Reserve
profile=fullfor rebuild/reindex and wide investigation.
Tooling Policy
- Prefer MCP tools before broad local text search when task is semantic code understanding:
rag_searchfirst for recall.- Then
lsp_definition,lsp_references,lsp_hover,lsp_diagnostics.
- Daily coding order:
- Symbol/API-level navigation: LSP (
lsp_*) first. - Broad cross-file context:
rag_search. - If embeddings are degraded: keep working in lexical mode temporarily, with diagnostics enabled.
- Symbol/API-level navigation: LSP (
- If MCP is down, recover via canonical preflight, then retry MCP tools.
Failure Handling
mcp:diagnosefailing:- Recheck PM2 status (
npm run daemon:status) - Re-run
npm run daemon:restart - Re-run
npm run mcp:diagnose
- Recheck PM2 status (
- RAG health failing:
- Confirm
OLLAMA_LOCAL_BASE_URL - Run
npm run rag:rebuild:zero
- Confirm
- If Ollama/model remains unavailable:
- Continue in lexical mode (non-blocking fallback).
- Mark risk explicitly in tracker/snapshot (
rag degraded, confidence reduced). - Keep MCP/LSP diagnostics active.
- Long rebuild:
- Keep process running; full profile can take minutes.
- If markdown volume is delaying index, run with
--docs-mode excludeand process docs in a second pass.
Command Reference
npm run daemon:startnpm run daemon:statusnpm run mcp:diagnosenpm run rag:rebuild:zeronpm run rag:rebuild:zero -- --docs-mode excludenpm run rag:health -- --jsonnpm run rag:index -- --profile full --docs-mode only
Done Criteria
- PM2 shows
agente-gpt,dashboard-web,chrome-proxyonline. http://localhost:3008/api/mcpresponds.mcp:diagnosepasses.- RAG rebuild completes with exit code
0, or fallback lexical explicitly registrado com risco.
When not to use it
- →When the task is solely about contract governance or runbook observability
- →When the task does not involve RAG, MCP, or LSP operations in this repository
- →When the task is about general coding without reliance on tool context
Limitations
- →The skill is specific to RAG/MCP/LSP operations in this repository
- →The skill's failure handling focuses on specific `npm run` commands
- →The skill assumes the presence of Ollama and PM2
How it compares
This skill provides a canonical, structured workflow for managing RAG, MCP, and LSP operations, including specific diagnostic and recovery steps, which is more complete than ad-hoc checks.
Compared to similar skills
rag-mcp-lsp-ops side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| rag-mcp-lsp-ops (this skill) | 0 | 5mo | Review | Intermediate |
| genkit-production-expert | 0 | 1mo | Review | Advanced |
| chaos-scenario | 0 | 2mo | Review | Advanced |
| reasoningbank-with-agentdb | 5 | 10mo | Review | Intermediate |
Try saying
Example prompts that trigger this skill in your AI assistant.
More by Ilenburg1993
View all by Ilenburg1993 →You might also like
genkit-production-expert
jeremylongshore
Build production Firebase Genkit applications including RAG systems, multi-step flows, and tool calling for Node.js/Python/Go. Deploy to Firebase Functions or Cloud Run with AI monitoring. Use when asked to "create genkit flow" or "implement RAG". Trigger with relevant phrases based on skill purpose.
chaos-scenario
petercort
Use when authoring, running, or reviewing chaos engineering experiments in this monorepo. Covers steady-state hypothesis, fault injection (service kill, latency, network partition, overload), result recording to CSV, and cleanup/restore. Triggers: "add chaos scenario", "new chaos test", "inject faul
reasoningbank-with-agentdb
ruvnet
Implement ReasoningBank adaptive learning with AgentDB's 150x faster vector database. Includes trajectory tracking, verdict judgment, memory distillation, and pattern recognition. Use when building self-learning agents, optimizing decision-making, or implementing experience replay systems.
agent-memory-mcp
davila7
A hybrid memory system that provides persistent, searchable knowledge management for AI agents (Architecture, Patterns, Decisions).
senior-ml-engineer
davila7
World-class ML engineering skill for productionizing ML models, MLOps, and building scalable ML systems. Expertise in PyTorch, TensorFlow, model deployment, feature stores, model monitoring, and ML infrastructure. Includes LLM integration, fine-tuning, RAG systems, and agentic AI. Use when deploying ML models, building ML platforms, implementing MLOps, or integrating LLMs into production systems.
ai-engineer
sickn33
Build production-ready LLM applications, advanced RAG systems, and intelligent agents. Implements vector search, multimodal AI, agent orchestration, and enterprise AI integrations. Use PROACTIVELY for LLM features, chatbots, AI agents, or AI-powered applications.