codinfy-agent-monitor
Provides real-time monitoring of AI agents, context, usage limits, workflows, files, Git, tests, builds, errors, secrets, and model-economy advice for tools like Claude Code, Codex, and Cursor, with 30 MCP tools for comprehensive observability.
README
<div align="center"> <picture> <source media="(prefers-color-scheme: dark)" srcset="./assets/brand/codinfy-logo-light.svg"> <source media="(prefers-color-scheme: light)" srcset="./assets/brand/codinfy-logo-dark.svg"> <img src="./assets/brand/codinfy-logo-dark.svg" width="410" alt="Codinfy"> </picture>
<br><br>
Codinfy Agent Monitor
Real-time mission control for AI engineering teams.
Observe agents, context, limits, workflows, Git, security and model economy from one local-first command center.
<br>
<br>
Overview · Quick start · Command center · MCP API · Architecture · Security · Contributing · License </div>
<br>
<p align="center"> <img src="./assets/readme/hero-glass.svg" width="100%" alt="Animated Codinfy Agent Monitor glass mission-control dashboard"> </p>
✦ The mission
AI agents move fast. Their context, tool calls, file changes, limits and failures usually remain scattered across terminals and provider interfaces. Codinfy Agent Monitor turns that invisible activity into a single operational picture.
| Mission system | What it reveals | |
|---|---|---|
| 🛰️ | Agent Radar | Active, idle, thinking, running, blocked, failed and completed agents |
| ◫ | Context & Limits | Context, current rate, daily and weekly usage with official/estimated provenance |
| ⟁ | Workflow Pulse | Tasks, owners, progress, blockers, files and a real-time event timeline |
| ◈ | AI Credit Saver | Model Need Score, safer model category and estimated token/cost savings |
| ⛨ | Safe Guard | Secret scanning, sensitive-file detection and review before commit |
| ⎇ | Dev Operations | Git state, tests, builds, environment and public-ready checks |
| ◉ | Codix Observer | Context-aware recommendations for risk, economy and next actions |
| ⇄ | Universal MCP | Claude Code, Codex, Cursor, Windsurf and other MCP-compatible hosts |
[!IMPORTANT] Exact provider limits are displayed as
officialonly when a host or adapter supplies them directly. Otherwise the interface clearly enables estimate mode. Codinfy Agent Monitor never presents a local estimate as provider truth.
⚡ Launch in 60 seconds
<details open> <summary><strong>01 — Install from source</strong></summary>
git clone https://github.com/bakalagoin/codinfy-agent-monitor.git
cd codinfy-agent-monitor
corepack enable
pnpm install
pnpm build
pnpm --filter codinfy-agent-monitor link --global
</details>
<details open> <summary><strong>02 — Initialize the local command center</strong></summary>
codinfy-agent-monitor init
codinfy-agent-monitor status
Your private operational data is created inside .codinfy-agent-monitor/ and excluded from Git.
</details>
<details> <summary><strong>03 — Open the animated TUI or local glass dashboard</strong></summary>
# Animated terminal mission control
codinfy-agent-monitor watch
# Local Fastify + WebSocket dashboard
codinfy-agent-monitor web
Open http://localhost:3579/dashboard. The server binds to 127.0.0.1 by default.
</details>
Without a global link:
pnpm codinfy status
pnpm codinfy watch
pnpm mcp
◫ Command center
╭────────────────── Codinfy Agent Monitor ──────────────────╮
│ Project: codinfy-agent-monitor │
│ Active AI: Codex · Command: /codinfy │
│ MCP: codinfy-agent-monitor │
├────────────────────────────────────────────────────────────┤
│ Context used ██████████████░░░░░░ 68% official │
│ Current rate ████████░░░░░░░░░░░░ 42% estimated │
│ Daily limit ████████████████░░░░ 81% estimated │
│ Weekly limit ████████░░░░░░░░░░░░ 42% estimated │
├────────────── AI Credit Saver & Smart Router ──────────────┤
│ Current model: Premium Reasoning │
│ Recommended: Standard Code · estimated saving: 45% │
│ Model switch: confirmation required │
├────────────────────────────────────────────────────────────┤
│ Agents: 5 active · 3 idle · Workflow: 7 tasks │
│ ✓ secrets.clean ✓ tests.passed ✓ build.passed │
╰────────────────────────────────────────────────────────────╯
© CODINFY PLATFORMS SASU · codinfy.com
<div align="center">
| Interface | Command | Experience |
|---|---|---|
| Snapshot | codinfy-agent-monitor status |
Fast operational overview |
| Animated TUI | codinfy-agent-monitor watch |
Live terminal bars and agent activity |
| Web glass UI | codinfy-agent-monitor web |
Responsive local dashboard + WebSocket |
| MCP server | codinfy-agent-monitor mcp |
Tool API for compatible AI environments |
| Slash command | /codinfy |
Native host-oriented monitoring flow |
</div>
⌘ Operator commands
<details open> <summary><strong>Monitoring & workflow</strong></summary>
codinfy-agent-monitor status
codinfy-agent-monitor agents
codinfy-agent-monitor context
codinfy-agent-monitor limits
codinfy-agent-monitor workflow
codinfy-agent-monitor tasks
codinfy-agent-monitor timeline
codinfy-agent-monitor files
codinfy-agent-monitor health
</details>
<details> <summary><strong>Engineering, Git & release safety</strong></summary>
codinfy-agent-monitor git
codinfy-agent-monitor diff
codinfy-agent-monitor tests --run
codinfy-agent-monitor build --run
codinfy-agent-monitor secrets
codinfy-agent-monitor check-command "git push --force"
codinfy-agent-monitor attribution-check
codinfy-agent-monitor review
codinfy-agent-monitor public-ready
codinfy-agent-monitor commit-message
codinfy-agent-monitor pr
codinfy-agent-monitor docs-check
codinfy-agent-monitor handoff
</details>
<details> <summary><strong>AI Credit Saver & Smart Model Router</strong></summary>
codinfy-agent-monitor saver
codinfy-agent-monitor budget
codinfy-agent-monitor cost
codinfy-agent-monitor model-score "audit the authentication architecture"
codinfy-agent-monitor model-advice "update the README"
codinfy-agent-monitor model-rules
codinfy-agent-monitor switch-model standard_code
codinfy-agent-monitor economy-plan
The router recommends configurable categories — fast_cheap, standard_code, advanced_code or premium_reasoning — instead of hard-coding provider model names. It never changes the active model without confirmation.
</details>
<details> <summary><strong>Beginner & guided mode</strong></summary>
codinfy-agent-monitor beginner
codinfy-agent-monitor next
codinfy-agent-monitor checklist
codinfy-agent-monitor glossary
codinfy-agent-monitor explain-error
codinfy-agent-monitor simple-report
codinfy-agent-monitor install-guide
codinfy-agent-monitor github-guide
codinfy-agent-monitor learn
</details>
<details> <summary><strong>Language, user level & environment</strong></summary>
codinfy-agent-monitor language auto
codinfy-agent-monitor language fr
codinfy-agent-monitor level beginner
codinfy-agent-monitor level expert
codinfy-agent-monitor safe on
codinfy-agent-monitor environment
codinfy-agent-monitor doctor
</details>
⬡ MCP Tool API
Start the local stdio server:
codinfy-agent-monitor mcp
{
"mcpServers": {
"codinfy-agent-monitor": {
"command": "codinfy-agent-monitor",
"args": ["mcp"]
}
}
}
| Tool domain | MCP tools |
|---|---|
| Status | monitor.status, monitor.open_dashboard, monitor.get_attribution |
| Agents | monitor.list_agents, monitor.register_agent, monitor.update_agent_state |
| Usage | monitor.get_context_usage, monitor.get_rate_limit_status, monitor.get_daily_usage, monitor.get_weekly_usage |
| Model economy | monitor.get_model_advice, monitor.get_model_score, monitor.get_budget_status, monitor.get_economy_plan |
| Workflow | monitor.create_task, monitor.update_task, monitor.list_tasks, monitor.get_workflow, monitor.timeline |
| Engineering | monitor.git_status, monitor.git_diff, monitor.test_status, monitor.build_status, monitor.environment_status, monitor.check_command |
| Release | monitor.commit_message, monitor.pr_summary, monitor.docs_check, monitor.handoff |
| Guided | monitor.simple_report, monitor.explain_error, monitor.model_rules, monitor.switch_model |
| Trust | monitor.scan_secrets, monitor.review_before_commit, monitor.alerts, monitor.recommendations |
| Reports | monitor.export_report, monitor.get_cost_estimate |
The integration test launches the built stdio server, connects an MCP client, lists tools and calls monitor.get_attribution.
⇄ AI environment adapters
<table> <tr> <td align="center"><strong>Claude Code</strong><br><sub>slash command + safe hook + MCP</sub></td> <td align="center"><strong>Codex</strong><br><sub>config.toml + AGENTS.md + MCP</sub></td> <td align="center"><strong>Cursor</strong><br><sub>mcp.json + rule + skill</sub></td> <td align="center"><strong>Windsurf</strong><br><sub>MCP config + /codinfy rule</sub></td> </tr> </table>
Ready-to-copy templates live under templates/. If a host does not expose custom slash commands, the CLI, TUI and local web dashboard remain universal fallbacks.
⎔ Architecture
flowchart LR
subgraph Hosts[AI environments]
Claude[Claude Code]
Codex[Codex]
Cursor[Cursor]
Windsurf[Windsurf]
end
subgraph Interfaces[Codinfy interfaces]
MCP[MCP stdio · 40 tools]
CLI[CLI · /codinfy]
TUI[Animated Ink TUI]
WEB[Fastify + WebSocket]
end
CORE[Core monitoring engine]
DB[(Local SQLite · WAL)]
OPS[Git · Tests · Build · Environment]
TRUST[Secret Scanner · Safe Guard · Review]
ROUTER[AI Credit Saver · Model Router]
Claude & Codex & Cursor & Windsurf --> MCP
MCP & CLI & TUI & WEB --> CORE
CORE --> DB
CORE --> OPS
CORE --> TRUST
CORE --> ROUTER
packages/core domain, SQLite, Git, security, model router, reports
packages/cli codinfy-agent-monitor + codinfy-monitor binaries
packages/tui animated Ink terminal mission control
packages/server local Fastify + WebSocket glass dashboard
packages/mcp-server stdio MCP server codinfy-agent-monitor
packages/adapter-* Claude Code, Codex, Cursor and Windsurf metadata
templates/ ready-to-copy /codinfy and MCP integrations
See the full architecture guide.
⛨ Trust center
| Control | Default |
|---|---|
| Data location | Local .codinfy-agent-monitor/ directory |
| Dashboard exposure | Loopback only — 127.0.0.1 |
| Secret output | Redacted before CLI, MCP and report output |
| Project commands | Tests/build run only with explicit --run |
| Model switching | Never automatic; confirmation is mandatory |
| Git behavior | Read-only monitoring; no implicit commit or push |
| Public-ready gate | Secrets + attribution + tests + build + sensitive files |
[!CAUTION] Never paste API keys, tokens, passwords,
.envcontent, private logs or production credentials into a public issue. Use private vulnerability reporting.
Read the Security Policy · Read the threat model
▣ Requirements
| Runtime | Requirement |
|---|---|
| Node.js | >= 22.13.0 for the built-in node:sqlite local store |
| pnpm | >= 10 |
| Git | Recommended for file and repository monitoring |
| Platforms | Windows, macOS and Linux by design |
◇ Build with us
corepack pnpm install
pnpm check
pnpm check runs lint, 15 tests — including a live MCP stdio client/server test — TypeScript build and formatting validation.
| Guide | Purpose |
|---|---|
| CONTRIBUTING.md | Local setup, branch flow and PR checklist |
| SECURITY.md | Private disclosure and support policy |
| LICENSE | Codinfy Attribution License 1.0 |
| CHANGELOG.md | Release history |
| Documentation | Installation, usage, MCP and security |
◉ Local storage
<details> <summary><strong>Open the private data layout</strong></summary>
.codinfy-agent-monitor/
├── config.json
├── metrics.sqlite
├── sessions/
├── agents/
├── workflows/
├── logs/
├── reports/
└── cache/
The entire directory is ignored by Git. .env*, credentials, logs, build output and local reports are never intended for publication.
</details>
⚖ License & attribution
This public repository is distributed under the custom Codinfy Agent Monitor Attribution License 1.0. Because it protects product identity and carries enhanced attribution conditions, it may be classified as source-available rather than OSI open source. Qualified legal review is recommended before high-stakes reliance.
The following identity must remain visible in copies, forks, distributions, interfaces, reports, exports and modified versions:
Codinfy Agent Monitor
/codinfy
codinfy-agent-monitor
© CODINFY PLATFORMS SASU
codinfy.com
Created by CODINFY PLATFORMS SASU
Bakala Goin — Founder & CEO
<div align="center"> <picture> <source media="(prefers-color-scheme: dark)" srcset="./assets/brand/codinfy-logo-light.svg"> <source media="(prefers-color-scheme: light)" srcset="./assets/brand/codinfy-logo-dark.svg"> <img src="./assets/brand/codinfy-logo-dark.svg" width="250" alt="Codinfy"> </picture>
Connect with Codinfy
Bakala Goin — Founder & CEO
<br><br>
Created by CODINFY PLATFORMS SASU<br> Bakala Goin — Founder & CEO<br> codinfy.com
<br>
© CODINFY PLATFORMS SASU · codinfy.com </div>
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.