BPFtraceSAFE
MCP server: using eBPF to tracing your kernel
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
A minimal MCP (Model Context Protocol) server that provides AI assistants with access to bpftrace kernel tracing capabilities.
Now implemented in Rust using the rmcp crate for better performance and type safety. The Python implementation is still available in the git history.
Features
- AI-Powered Kernel Debugging: Enable AI assistants to help you debug complex Linux kernel issues through natural language - no eBPF expertise required
- Discover System Trace Points: Browse and search through thousands of kernel probes to find exactly what you need to monitor - from system calls to network packets
- Rich Context and Examples: Access a curated collection of production-ready bpftrace scripts for common debugging scenarios like performance bottlenecks, security monitoring, and system troubleshooting
- Secure Execution Model: Run kernel traces safely without giving AI direct root access - MCPtrace acts as a secure gateway with proper authentication
- Asynchronous Operation: Start long-running traces and retrieve results later - perfect for monitoring production issues that occur intermittently
- System Capability Detection: Automatically discover what tracing features your kernel supports, including available helpers, map types, and probe types
Why MCPtrace?
Debugging kernel issues traditionally requires deep eBPF expertise. MCPtrace changes that.
By bridging AI assistants with bpftrace (the perfect eBPF tracing language), MCPtrace lets you debug complex system issues through natural conversation. Just describe what you want to observe - "show me which processes are opening files" or "trace slow disk operations" - and let AI generate the appropriate kernel traces.
AI never gets root access. MCPtrace acts as a secure gateway, and with its rich collection of example scripts and probe information, AI has everything n
b6e6dae686e1OBSERVED · 2026-10-07Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add mcptrace --env OPENAI_API_KEY=${OPENAI_API_KEY} -- uvx mcptrace{
"mcpServers": {
"mcptrace": {
"command": "uvx",
"args": [
"mcptrace"
],
"env": {
"OPENAI_API_KEY": "${OPENAI_API_KEY}"
}
}
}
}Exposed tools (4)
3 read · 1 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
exec_program | write | |
get_result | read | |
list_helpers | read | |
list_probes | read |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- none-observed
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (5)
fastmcp
> ⚠️ **WARNING**: This MCP server executes system-level tracing commands with elevated privileges. Proper security configuration is CRITICAL.
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
curl --proto '=https' --tlsv1.2 -sSf https://sh.rustup.rs | sh
doc/compressed_output.gif
Gates applied: no_behavioural_pass.
b6e6dae686e1full audit observations/trust-audit/mcp-server/eunomia-bpf__bpftrace.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-07 | b6e6dae686e1 | SAFE | B | 89 | first audit |
Questions
What is the BPFtrace MCP server?
MCP server: using eBPF to tracing your kernel
What tools does BPFtrace expose?
4 in total: 3 read-only, 1 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is BPFtrace safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B.
What credentials does BPFtrace need?
It reads OPENAI_API_KEY from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does BPFtrace run?
It speaks stdio, so it runs as a local process your client starts. It is published on PyPI as mcptrace.
How current is this page?
The grade is for one exact copy of the source (b6e6dae686e1), read on 2026-10-07. The repository is watched and re-audited when it changes.