Orygn Opa MCP Server

@orygn/opa-mcp npm v0.6.0 Source verified

Published by orygnscode — publish provenance cryptographically ties this package to that repository. That is proof of origin, not an official vendor package.

Model Context Protocol server for Open Policy Agent (OPA). Author, validate, debug, and explain Rego policies through any MCP-compatible client.

View on npm ↗ Repository ↗ Raw report (JSON) ↗

Trust grade

94/100

Last scanned 2026-09-09 get badge →

Trust

A · 94/100

Adoption risk for you: the threat score, then adjusted down for blast radius, publisher verification and how much the scan could see. Deterministic; every point is auditable.

Capability

High

Blast radius if it went rogue — what the server’s tools could reach. Independent of trust.

Coverage

Source

How much the scan could actually inspect. Shallow coverage is stated, never hidden.

A Why this grade threat 100 − adoption risk = 94/100

The grade answers one question — how safe is this server for you to adopt — so it is computed in two auditable stages. Nothing below is an opinion or an LLM's guess; every line is a real term the deterministic engine applied, and the same input always yields the same number.

1. Threat score — 100 − 0 = 100. What the published surface and source actually contain:

The deterministic scan raised no scored threat in the surface it inspected — the threat score stayed at 100. Capability observations and advisory notes are recorded but never lower it.

2. Client adoption risk — 100 − 6 = 94. Three small, subtract-only factors that reflect your risk in adopting it — a clean scan proves less on a powerful, unverified or barely-inspectable package, so the grade says so plainly:

Points	Adoption-risk factor
−6	capability blast radius (high) — client exposure if the model is manipulated

Capability observations and info notes are shown under Findings but never scored. Open any row's finding below for the file, line and evidence behind a deduction.

Findings 17

high Dynamic evaluation of a non-literal value (dist/lib/opa-cli.js)MTC-SRC-010

In the server's implementation (`dist/lib/opa-cli.js:196`): Evaluating a runtime value as code (rather than a fixed literal) executes whatever reaches it — a direct RCE primitive, and almost never necessary in legitimate code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: ons). */ async eval(input, signal) { // Inline source becomes a temp file added to --data. if (

Fix: Review this call path: confirm it never receives unsanitized tool input, constrain it, or remove it. Treat a server whose code reaches these sinks as high-capability regardless of what its tools claim.

Location: server dist/lib/opa-cli.js

high Tool "rego_eval" exposes command/code executionMTC-CAP-001

Tool "rego_eval" appears to run shell commands or evaluate code (keyword "eval" in tool name). Arbitrary execution driven by model input is one of the most dangerous MCP capabilities; combined with any untrusted input it becomes RCE.

Fix: Sandbox execution, allowlist commands/arguments, and never pass model output to a shell unescaped.

Location: tool rego_eval

high Tool "rego_eval_with_explain" exposes command/code executionMTC-CAP-001

Tool "rego_eval_with_explain" appears to run shell commands or evaluate code (keyword "eval" in tool name). Arbitrary execution driven by model input is one of the most dangerous MCP capabilities; combined with any untrusted input it becomes RCE.

Fix: Sandbox execution, allowlist commands/arguments, and never pass model output to a shell unescaped.

Location: tool rego_eval_with_explain

high Tool "rego_eval_with_profile" exposes command/code executionMTC-CAP-001

Tool "rego_eval_with_profile" appears to run shell commands or evaluate code (keyword "eval" in tool name). Arbitrary execution driven by model input is one of the most dangerous MCP capabilities; combined with any untrusted input it becomes RCE.

Fix: Sandbox execution, allowlist commands/arguments, and never pass model output to a shell unescaped.

Location: tool rego_eval_with_profile

high Tool "rego_eval_with_coverage" exposes command/code executionMTC-CAP-001

Tool "rego_eval_with_coverage" appears to run shell commands or evaluate code (keyword "eval" in tool name). Arbitrary execution driven by model input is one of the most dangerous MCP capabilities; combined with any untrusted input it becomes RCE.

Fix: Sandbox execution, allowlist commands/arguments, and never pass model output to a shell unescaped.

Location: tool rego_eval_with_coverage

high Tool "opa_exec" exposes command/code executionMTC-CAP-001

Tool "opa_exec" appears to run shell commands or evaluate code (keyword "exec" in tool name). Arbitrary execution driven by model input is one of the most dangerous MCP capabilities; combined with any untrusted input it becomes RCE.

Fix: Sandbox execution, allowlist commands/arguments, and never pass model output to a shell unescaped.

Location: tool opa_exec

high Dynamic code execution in server code (dist/lib/opa-cli.js)MTC-SRC-001

In the server's implementation (`dist/lib/opa-cli.js:196`): Evaluating strings as code is the most direct RCE primitive; if any tool input reaches it, the server executes attacker-chosen code. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: ons). */ async eval(input, signal) { // Inline source becomes a temp file added to --data. if (

Location: server dist/lib/opa-cli.js

high Shell/command execution in server code (dist/lib/opa-cli.js)MTC-SRC-002

In the server's implementation (`dist/lib/opa-cli.js:390`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: file. */ async exec(input, signal) { if (input.inputPaths.length === 0) { throw new Error('

Location: server dist/lib/opa-cli.js

high Shell/command execution in server code (dist/lib/subprocess.js)MTC-SRC-002

In the server's implementation (`dist/lib/subprocess.js:10`): Spawning a shell/process is command-execution capability; with unsanitized tool input it is command injection / RCE. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: rt { spawn } from 'node:child_process'; import { buildChildEnv } from './child-env.js'; /** * Per-stream capture ceilin

Location: server dist/lib/subprocess.js

medium Untrusted input can drive an external actionMTC-FLOW-005

Untrusted-input tools ([conftest_pull, opa_get_policy]) co-exist with external-action tools ([rego_eval, rego_eval_with_explain, rego_eval_with_profile, rego_eval_with_coverage, opa_exec]). A prompt injection could cause unwanted external actions, though no direct sensitive-data leak path was found.

Evidence: untrusted [conftest_pull, opa_get_policy] → sinks [rego_eval, rego_eval_with_explain, rego_eval_with_profile, rego_eval_with_coverage, opa_exec]

Fix: Require confirmation for state-changing/egress actions triggered after processing untrusted content.

Location: flow conftest_pull → rego_eval

medium Hardcoded egress to an external endpoint (dist/server.js)MTC-SRC-003

In the server's implementation (`dist/server.js:41`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: 'u', id); await fetch(`https://opa-mcp-telemetry.gibbidaniel.workers.dev/ping?${params.toString()}`, {

Location: server dist/server.js

medium Hardcoded egress to an external endpoint (dist/tools/helpers/playground-share.js)MTC-SRC-003

In the server's implementation (`dist/tools/helpers/playground-share.js:89`): A hardcoded outbound call to a fixed external host inside server code is a classic exfiltration/telemetry channel — especially paired with reads of local data. This is read from the code itself — not from the tool description — so a poisoned server cannot hide it behind honest-looking metadata.

Evidence: response = await fetch('https://api.github.com/gists', { method: 'POST',

Location: server dist/tools/helpers/playground-share.js

low Mutating tool "rego_eval" declares no destructiveHintMTC-CAP-005

Tool "rego_eval" can mutate/egress but declares no destructiveHint. Clients that don't default to spec-safe behavior may not prompt before running it.