AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
MCP verified MIT Self-run

Revoke Probe

mcp-revoke-trust-network-revoke-probe · by Revoke-Trust-Network

Local-first reachability check for AI coding-agent and MCP workflows

No reviews yet
0 installs
13 views
0.0% view→install

Install

$ agentstack add mcp-revoke-trust-network-revoke-probe

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.0 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access Used
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets Used
  • Dynamic code execution No

From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/mcp-revoke-trust-network-revoke-probe)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
2mo ago

Declared compatibility

Claude CodeClaude DesktopCursorWindsurf

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Revoke Probe? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Revoke Probe

Local-first reachability check for AI coding-agent and MCP workflows.

[](https://opensource.org/licenses/MIT) []() []()

Revoke Probe helps you answer one narrow question:

> What sensitive-looking local surfaces could this developer workflow reach?

It scans for common credential-adjacent path patterns and can be paired with fake canaries to produce a sanitized report you can share after review.

No telemetry. No source upload. No real secret values in the report. No vendor-blame claims.

revoke-probe is the free local probe used by Agent Leak Bench.

Install

macOS

brew install Revoke-Trust-Network/tap/revoke-probe

Linux / WSL / Git Bash

curl -fsSL https://raw.githubusercontent.com/Revoke-Trust-Network/revoke-probe/v0.1.0/install.sh -o revoke-probe-install.sh
sh revoke-probe-install.sh

From source

git clone https://github.com/Revoke-Trust-Network/revoke-probe
cd revoke-probe
./revoke-probe --scan

Usage

$ revoke-probe --scan

Output is a structured report of which AI-tool-accessible paths match common high-value patterns, sorted by risk tier. The report is designed to be safe to share after review because it must not include real secret values.

For automation:

revoke-probe --scan --json
revoke-probe --scan --path /path/to/test/root

What it scans (path pattern catalog)

See [docs/path-patterns.md](docs/path-patterns.md) for the full list. TL;DR — 18 path families across SSH/cloud creds, dotenv variants, browser profiles, and framework config dirs.

Fake canaries

The benchmark uses fake values only:

  • fake AWS key marker
  • fake Stripe key marker
  • fake GitHub token marker
  • fake OpenAI key marker
  • fake SSH private key marker

Do not paste real credentials into issues, reports, screenshots, posts, or support emails.

What it proves

  • Which sensitive path families are reachable on this machine.
  • Which findings are worth reviewing before a team standardizes an AI coding workflow.

What it does NOT prove

  • It does not prove a vendor stole your data.
  • It does not upload anything anywhere.
  • It does not require real secret values.
  • It does not prevent agents from reading these files by itself.
  • It does not replace endpoint security, MDM, or secrets management.
  • It does not modify your filesystem outside controlled benchmark fixtures.

Why this exists

AI coding tools moved fast. Team policy did not. Most developers can name the AI assistant they installed, but they cannot quickly answer a more operational question:

> What can my current AI coding stack touch on this machine?

This benchmark exists to make that question visible, reproducible, and boring enough to discuss without vendor-blame theater.

Contributing

PRs welcome, especially:

  • Additional dotfile patterns we missed
  • Translations of the report output

Roadmap

  • [x] Path pattern catalog
  • [x] Shell CLI scanner
  • [x] JSON output mode
  • [x] Homebrew tap
  • [ ] eBPF-backed real-time read tracing (Linux)
  • [ ] CI integration mode

Pro tier — revoke-node

revoke-probe focuses on scan + report.

revoke-node is the commercial product line around the same problem space. This README intentionally avoids capability promises that are not verified inside this repo.

Team Pilot

Teams using Cursor, Claude Code, Codex CLI, Windsurf, VS Code extensions, or MCP servers can request a 30-day audit:

support@revokenode.io
Subject: Team Pilot

License

MIT © 2026 Revoke-Trust-Network

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.0 Imported from the upstream source.