Monitor
23 scored tools, ranked by overall score. Last checked on 25 October 2026. Refine with the Finder →
The three to remember
The go-to self-hostable solution, which remained MIT-licensed after the acquisition by ClickHouse.
- · Utility 9/10, best of the selection
- · Free tier to get started
- · No known security caveat
Run it before and after every community MCP installation.
- · Utility 8.5/10
- · Free, no limits
- · No known security caveat
The nmap of LLMs, to integrate into a red-teaming CI pipeline.
- · Utility 8/10
- · Free, no limits
- · No known security caveat
| Tool | Price | Score | Sec. |
|---|---|---|---|
| LangfuseRisingOpen source LLM observability: traces, costs, evaluations, prompt management. | Freemium | 9.3 | Security: nothing to flag |
| mcp-scanRisingScans your MCP servers and installed skills for tool poisoning, rug pulls, and injections. | Free | 9.3 | Security: nothing to flag |
| PromptfooRisingPrompt, agent, and RAG testing and red-teaming, CI-integrable. | Freemium | 9.0 | Security: nothing to flag |
| Arize PhoenixRisingLLM application tracing and evaluation, built on OpenTelemetry. | Freemium | 8.9 | Security: nothing to flag |
| garakRisingNVIDIA's LLM vulnerability scanner: jailbreaks, leaks, hallucinations, injections. | Free | 8.9 | Security: nothing to flag |
| Cisco mcp-scannerRisingOpen source scanner by Cisco based on YARA to analyze MCP tool descriptions and schemas. | Free | 8.7 | Security: nothing to flag |
| HeliconeRisingObservability proxy: costs, latency and logs of all your LLM requests by changing one URL. | Freemium | 8.7 | Security: nothing to flag |
| ToolHiveRisingRuns each MCP server in an isolated container with managed permissions and secrets. | Free | 8.7 | Security: nothing to flag |
| PortkeyRisingAI gateway with routing, guardrails, and observability, open source since March 2026. | Freemium | 8.6 | Security: nothing to flag |
| PyRITRisingMicrosoft's AI red-teaming toolkit. | Free | 8.3 | Security: nothing to flag |
| Approbation humaine (human-in-the-loop)Pauses the agent and requests confirmation before a sensitive action. | Free | 8.2 | Security: nothing to flag |
| Permissions (Claude Code)Allow, deny, and ask rules per tool and per pattern. | Free | 8.2 | Security: nothing to flag |
| LangSmithObservability and evaluation platform from LangChain, usable outside LangChain. | Freemium | 8.1 | Security: nothing to flag |
| Lakera GuardDetects direct and indirect prompt injections before the model or tool is called. | Paid | 8.0 | Security: nothing to flag |
| BraintrustEvaluation and logging platform for AI products, oriented towards product teams. | Freemium | 7.9 | Security: nothing to flag |
| StrixOpen-source AI agent pentesting: agents run your application, find vulnerabilities, and prove each one with a PoC. | Freemium | 7.7 | Security: one caveat, see the card |
| Prompt SecurityManaged guardrail for applications and agents. | Paid | 7.6 | Security: nothing to flag |
| LLM Guard (Protect AI)Watch outOpen source input and output scanners: injection, personal data, toxicity. | Free | 7.4 | Security: nothing to flag |
| E2B Code InterpreterRisingIsolated micro-VM for executing code produced by an agent. | Freemium | 7.3 | Security: one caveat, see the card |
| Invariant GuardrailsWatch outSecurity rules applied in real time to agent tool calls. | Freemium | 6.8 | Security: nothing to flag |
| MCP DefenderDesktop app that intercepts and filters MCP traffic from installed AI clients. | Freemium | 6.3 | Security: one caveat, see the card |
| DaytonaFast-starting sandboxes for agents. | Freemium | 6.1 | Security: one caveat, see the card |
| Secrets pour agents (1Password, Vault)Provide secrets to an agent without putting them in the prompt. | Paid | 6.1 | Security: one caveat, see the card |
App : A product you use directly · Tool : A precise action the AI can call · MCP : Standardises the connection between an AI application and systems that expose tools or data · Built-in mode : A feature to switch on inside an assistant you already use (ChatGPT, Claude, Gemini): nothing to install · Skill : A reusable method: instructions, procedures, sometimes scripts or templates · Plugin : A set of capabilities installed into an AI application or a host
Method
Overall score = utility (rated by the editorial team), reliability (GitHub activity, official publisher, reachable site), security (where data goes, permissions, known incidents) and accessibility (install, required level, price), weighted per category. A tool inactive for over six months is flagged and leaves the recommendations. No ranking is sold.
Missing a tool? A score looks wrong?
Write to the editors