Install
$ agentstack add mcp-agentopssec-mcp-radar ✓ scanned · ✓ verified — works with Claude Code, Cursor, and more.
Security review
✓ PassedNo 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 No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ 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.
About
MCP Radar
Trust scores for MCP servers.
MCP Radar is a public trust and discovery layer for MCP servers. It helps developers evaluate MCP servers based on permissions, maintenance, install method, package quality, documentation, and security risk before installing them.
Think of it as:
The trust index for MCP servers
Why This Exists
MCP servers are becoming a new supply chain for AI agents. Developers need to know whether a server is maintained, documented, permissioned reasonably, and safe enough to connect to a real agent workflow.
MCP Radar answers questions like:
- Who maintains this MCP server?
- What permissions does it require?
- Does it expose filesystem, shell, network, or database access?
- Is the package pinned and released responsibly?
- Is the repository active?
- Does the project have clear docs and schemas?
- Are there known vulnerabilities?
- Is the install method safe?
- How does it compare to alternatives?
MCP Radar turns those signals into transparent trust scores.
Install
npm install -g @agentopssec/mcp-radar
Or run it without installing:
npx -y @agentopssec/mcp-radar search github
Update
mcp-radar update # check the registry, prompt before installing
mcp-radar update --yes # update without prompting
Primary Workflow
MCP Radar starts with searchable server profiles and CLI score lookup:
mcp-radar score owner/repo
The workflow should do three things well:
- Show whether an MCP server is safe enough to install.
- Explain the signals behind the score.
- Make scores reusable through profiles, badges, and JSON.
CLI
mcp-radar search github
mcp-radar score owner/repo
mcp-radar compare github-server filesystem-server
mcp-radar badge owner/repo
mcp-radar submit ./mcp-server.json
mcp-radar score-doctor ./mcp-doctor-scan.json
mcp-radar validate
mcp-radar registry update --from registry.json
mcp-radar update [--yes]
Standalone and Stack Use
MCP Radar runs on its own using its curated local registry. Projects can also provide .mcp-radar/registry.json for local/private server profiles:
mcp-radar search github
mcp-radar score modelcontextprotocol/server-github
mcp-radar registry update --from ./registry.json
When used with the full AgentOpsSec stack, it can score servers found by MCP Doctor without importing MCP Doctor code:
mcp-doctor scan --json --output mcp-doctor-scan.json
mcp-radar score-doctor ./mcp-doctor-scan.json
What MCP Radar Scores
MCP Radar scores MCP servers using signals such as:
- Repository activity
- Maintainer activity
- Stars and forks
- License
- Package age
- Release history
- Known vulnerabilities
- Install method
- Required permissions
- Tool categories
- Filesystem access
- Shell access
- Network access
- Token handling
- Environment variables
- Docker support
- Documentation quality
- Schema quality
- Prompt injection risk indicators
Example Profile
MCP Radar Score by github.com/AgentOpsSec
Official GitHub MCP Server
Score: A-
Risk: Medium
Reason: Powerful repo access, but maintained and documented
Random Database MCP Server
Score: D
Risk: High
Reason: Broad database access, low maintainer activity, no version pinning, weak docs
Score Shape
{
"tool": {
"name": "MCP Radar",
"by": "github.com/AgentOpsSec",
"repository": "github.com/AgentOpsSec/mcp-radar"
},
"name": "example-mcp-server",
"score": "B-",
"risk": "medium",
"categories": ["github", "filesystem"],
"findings": [
{
"type": "permission",
"severity": "medium",
"message": "Requires repository write access"
},
{
"type": "maintenance",
"severity": "low",
"message": "Last release was 22 days ago"
}
]
}
Public Outputs
MCP Radar is designed to produce reusable public artifacts:
- MCP server profile pages
- Searchable registry entries
- Transparent scoring breakdowns
- JSON API responses
- Score badges
- Community submissions
- Comparison views
Design Principles
- Open scoring model
- Transparent findings
- Public by default
- Useful before installation
- Community-extensible registry
- Clear permission labels
- Security and maintenance signals together
- Easy integration with local scanners
Initial Release Scope
The initial release includes a public registry, curated MCP server profiles, transparent scoring, search, badges, JSON output, and community submissions.
1.0: Curated Registry
- Build a public MCP server registry
- Add an initial curated server list
- Create server profile pages
- Label tool categories and permissions
- Show install method and package metadata
- Include repository and license metadata
1.0: Scoring Model
- Define a transparent score model
- Score maintenance signals
- Score permission risk
- Score install safety
- Score documentation and schema quality
- Explain each finding behind the final score
1.0: Search, API, and Badges
- Add basic server search
- Add server comparison
- Generate score badges
- Expose public JSON score output
- Support community submissions
- Make scores consumable by local tools
Registry Validation
mcp-radar validate
mcp-radar validate ./registry.json
mcp-radar registry update --from ./registry.json
Registry updates are local to the current project and write .mcp-radar/registry.json.
Output
Reports use plain-language status words rather than raw exit codes:
ok— the step ran successfully (green).failed (exit N)— the step exited non-zero (red); the original code is preserved.skipped (reason)— the step was not applicable (dim).
Severity colors follow the AgentOpsSec palette (safe = green, warning = amber, risk = red). The palette honors NO_COLOR and FORCE_COLOR, and JSON / CSV output stays plain.
- Repo: https://github.com/AgentOpsSec/mcp-radar
- npm: https://www.npmjs.com/package/@agentopssec/mcp-radar
- AgentOpsSec stack: https://github.com/AgentOpsSec/stack
- Website: https://AgentOpsSec.com
Author
Created and developed by Aunt Gladys Nephew.
- Website: https://auntgladysnephew.com
- GitHub: https://github.com/auntgladysnephew
- X: https://x.com/AGNonX
Source & license
This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: AgentOpsSec
- Source: AgentOpsSec/mcp-radar
- License: MIT
- Homepage: https://AgentOpsSec.com
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet — be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.