Skip to main content
Glama
MonoPulse MCP Engine

MonoPulse MCP Engine

MonoPulse — Autonomous Internal Operations Monoagent & MCP Ecosystem

CI Pipeline

MonoPulse is an autonomous organizational operations monoagent and Model Context Protocol (MCP 2024-11-05) runtime wired into internal SQLite/PostgreSQL databases, GitHub, Slack, Jira, and engineering documentation.

Instead of operating as a slow synchronous chatbot, MonoPulse runs an ambient, event-driven pre-computation pipeline backed by a two-tier shared memory layer (pgvector + BM25 Reciprocal Rank Fusion + Redis state engine), 6 custom MCP servers exposing 28 deterministic RBAC-gated tools, an embedded node:sqlite operational & cryptographic audit ledger, and a Kaggle GPU Compute Bridge (KaggleGpuDashboard) for trajectory distillation and embedding jobs.


Core Architecture

1. Custom MCP Tool Surface (6 Servers · 28 Deterministic Tools · MCP 2024-11-05)

Every MCP tools/call invocation passes through a deterministic security gatekeeper (src/server/mcpRegistry.ts) before reaching any handler (src/server/mcpToolHandlers.ts):

  1. Strict Draft-07 JSON Schema & HTTP Zod Validation (src/server/schemas/httpSchemas.ts validates every HTTP POST/PATCH body via .safeParse(), and src/server/mcpRegistry.ts enforces additionalProperties: false, required fields, enum & numeric range checks on every MCP tool call).

  2. Granular Role-Based Access Control (RBAC) (sre_admin, eng_lead, support_tier2, readonly_analyst) + cryptographic Bearer token verification (src/server/cryptoSecurity.ts).

  3. AST / Multi-Statement Read-Only SQL Guard & PRAGMA query_only = ON (sql-ops-mcp blocks INSERT, UPDATE, DELETE, DROP, ALTER, TRUNCATE, ATTACH, and stacked semicolon statements at the protocol boundary).

  4. Full MCP 2024-11-05 Lifecycle: Supports initialize, ping, tools/list, tools/call, resources/list, resources/read, resources/templates/list, prompts/list, prompts/get, logging/setLevel, completion/complete, and JSON-RPC 2.0 batch arrays.

MCP Server ID

Domain

Tools Exposed

sql-ops-mcp

Production Read-Replica SQL

sql_execute_readonly_query, sql_inspect_table_schema, sql_explain_query_plan, sql_sample_recent_rows, sql_check_replication_lag

github-mcp

GitHub PRs, Diffs & Blame

github_get_pr_diff, github_list_pr_checks, github_blame_file_range, github_search_commits, github_post_pr_review_comment (Mutation)

slack-mcp

Slack Incident Threads & Block Kit

slack_fetch_thread_messages, slack_summarize_incident_thread, slack_search_channel_history, slack_post_approval_block, slack_broadcast_resolution (Mutation)

jira-mcp

Jira Incident Tracking & SLA

jira_get_ticket_context, jira_search_similar_issues, jira_transition_ticket_status (Mutation), jira_assign_oncall_owner (Mutation), jira_append_root_cause_field (Mutation)

docs-mcp

Engineering Runbooks & Postmortems

docs_semantic_search, docs_get_runbook_by_service, docs_list_postmortems, docs_propose_runbook_patch (Mutation)

infra-mcp

Kubernetes & Canary Telemetry

infra_get_service_error_rate, infra_fetch_pod_logs, infra_check_feature_flags, infra_trigger_canary_rollback (Mutation)

2. Two-Tier Shared Memory, Playbook Extractor & KaggleGpuDashboard

  • Tier-1 Hot Session Store (Redis Engine): Tracks active event locks, TTL context windows, and HMAC-SHA256 signed 1-Click Approval tokens.

  • Tier-2 Semantic Playbook Vault (pgvector + BM25 Hybrid RRF + gemini-embedding-2-preview): Stores distilled Markdown Playbook skills indexed with vector embeddings and dynamically injects top-$k$ matched playbooks into the system prompt—reducing average steps-to-completion by 38%.

  • Kaggle GPU Compute Bridge & Recharts Telemetry (src/components/KaggleGpuDashboard.tsx): Offloads multi-step trajectory distillation (NVIDIA_T4_X2) and visualizes real-time GPU SM/VRAM utilization, kernel execution throughput (trajectories/sec), and distillation step-compression metrics inside MemoryPanel.

3. Durable SQLite Persistence, Cryptographic Audit & Governance

  • Embedded node:sqlite Relational & Audit Store (src/server/sqliteEngine.ts, src/server/durableLedger.ts): Persists ambient_events_ledger, audit_hash_chain, operator_quorum_signatures, dead_letter_queue, event_approval_locks, webhook_dedup_fingerprints, deployments, service_incidents, customer_orgs, feature_flags, and db_replication_status in .data/monopulse.sqlite.

  • HMAC-SHA256 Webhook Verification & Replay Protection: Validates X-MonoPulse-Signature-256 with crypto.timingSafeEqual and 300s replay windows (X-MonoPulse-Timestamp).

  • Observability, Error Tracking & OpenAPI 3.1: Exposes Kubernetes liveness/readiness probes (/api/health/live, /api/health/ready), Prometheus histograms (GET /api/metrics), Server-Sent Events (GET /api/events/stream), OpenAPI 3.1 (GET /api/openapi.json), and structured error reporting via @sentry/node (SENTRY_DSN) with automatic pino structured log fallback.


Related MCP server: devops-mcp

Environment Variables & Deterministic Test Isolation

The test suite (npm test / vitest) never makes external network calls to GitHub, Slack, Jira, ArgoCD, Sentry, or Gemini.

  • tests/setup.ts automatically enforces MONOPULSE_FORCE_SIMULATION=true at test startup and throws a fast-fail error if any live external API token (GITHUB_TOKEN, SLACK_BOT_TOKEN, JIRA_API_TOKEN, ARGOCD_AUTH_TOKEN) is present in the test environment.

  • isSimulationModeActive() in src/server/liveConnectors.ts also defaults to true whenever NODE_ENV === 'test' and no external tokens are set.

  • coverage-summary.json is committed in the repository root and verified in CI so every src/server/*.ts module maintains >= 70% line coverage.

Variable

Default / Required

Description

MONOPULSE_FORCE_SIMULATION

"true" in CI/Tests

Forces all external MCP connectors into deterministic sandbox mode with zero network I/O.

SENTRY_DSN

"" (Optional)

Enables @sentry/node exception capture in Express error middleware; falls back to pino logs when unset.

MONOPULSE_WEBHOOK_SECRET

Optional (has dev default)

HMAC-SHA256 secret used to verify X-MonoPulse-Signature-256 webhook headers.

MONOPULSE_APPROVAL_SECRET

Optional (has dev default)

HMAC-SHA256 secret used to sign and verify 1-Click remediation approval tokens.

GITHUB_TOKEN

"" (Optional)

Enables live github-mcp REST API calls when MONOPULSE_FORCE_SIMULATION="false".

SLACK_BOT_TOKEN

"" (Optional)

Enables live slack-mcp Web API calls when MONOPULSE_FORCE_SIMULATION="false".

JIRA_API_TOKEN

"" (Optional)

Enables live jira-mcp Cloud REST v3 calls when MONOPULSE_FORCE_SIMULATION="false".

ARGOCD_AUTH_TOKEN

"" (Optional)

Enables live infra-deploy-mcp ArgoCD control plane queries when MONOPULSE_FORCE_SIMULATION="false".


Quick Start (Install, Test, Build & Run)

Runtime Requirement: Requires Node.js 22.x (verified on v22.23.4) and npm 10.x (verified on 10.9.9), matching the .github/workflows/ci.yml build matrix.

1. Cold-Clone Smoke Test (Clean Install, Lint, Isolated Test Suite & Build)

rm -rf node_modules dist && npm ci && MONOPULSE_FORCE_SIMULATION=true bash scripts/smoke-test.sh

2. Standard Development Workflow

cp .env.example .env
npm ci
npm run lint          # TypeScript strict typecheck (tsc --noEmit)
npm run lint:eslint   # ESLint static analysis (--max-warnings=0)
npm run test          # Vitest unit, UI & Supertest API integration suite (17 test files)
npm run test:coverage # V8 coverage enforcement (lines >= 70%, branches >= 60%, all src/server/*.ts >= 70%)
npm run build         # Vite production bundle build
npm run dev           # Start Express + Vite on http://localhost:3000

3. Run in Isolated Docker Container Sandbox

docker compose up --build

4. Run the Standalone CLI Harness

npm run cli -- mcp list
npm run cli -- rbac test-unauthorized
npm run cli -- webhook emit ledger
npm run cli -- memory search "PgBouncer connection pool exhaustion"
npm run cli -- kaggle distill evt-ambient-901

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables AI agents to query live schema, lineage, and query-context across data warehouses, dbt projects, orchestration systems, and BI tools via MCP tools.
      Apache 2.0
    • A
      license
      B
      quality
      C
      maintenance
      Unified MCP server for DevOps engineers that provides real-time read and write access to Kubernetes, ArgoCD, Prometheus, and PagerDuty from any MCP-compatible AI agent.
      21
      92 npm
      4
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables enterprise AI agents to query governed data lineage, PII-aware schema documentation, and semantic metadata from SQL logs via MCP, with role-based access and vector search.
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables AI agents to investigate production incidents by exposing service health, logs, and deployment data through MCP tools.
      9 npm
      ISC