mcp-opa-authz
io.github.kanywst/mcp-opa-authz
About mcp-opa-authz
mcp-opa-authz is an MCP server published by the GitHub account kanywst. Authorization answers from real policy code: evaluate Rego locally, or ask an AuthZEN 1.0 PDP.
It runs on your own machine: your AI client starts it from the Docker (OCI) package ghcr.io/kanywst/mcp-opa-authz:0.6.2, with no version pinned.
It asks for one secret: AUTHZEN_PDP_TOKEN. Give it a key with the least access that still works.
Version 0.6.2 was published on 25 October 2026, and the source code is public, so you can read it before installing.
Written from the registry entry. 7 automated checks on registry metadata: 4 pass, 0 warn, 0 fail, 3 info. How we check.
What it asks for
| Variable | Required | Secret | Description |
|---|---|---|---|
AUTHZEN_PDP_URL | No | No | Default AuthZEN Access Evaluation endpoint, e.g. https://pdp.example.com/access/v1/evaluation. Only evaluate_policy works without it. |
AUTHZEN_PDP_TOKEN | No | Yes | Authorization header value for the PDP. A value with no scheme is sent as "Bearer <token>". |
AUTHZEN_PDP_TIMEOUT | No | No | Per-request timeout for PDP calls, as a Go duration. Default 10s. |
MCP_OPA_EVAL_TIMEOUT | No | No | Wall-clock limit on a single Rego evaluation, as a Go duration. Default 5s. |
MCP_OPA_ALLOW_NETWORK_BUILTINS | No | No | Re-enable http.send, net.lookup_ip_addr and opa.runtime inside evaluated policies. Off by default: policy source evaluated here comes from a model and runs inside the server process. |
Checks
| Check | Result | Detail |
|---|---|---|
| Source repository linked | Pass | https://github.com/kanywst/mcp-opa-authz |
| Version pinned | Pass | Every package has an explicit version |
| Secrets requested | Info | Asks for 1 secret: AUTHZEN_PDP_TOKEN |
| Known package registry | Pass | All packages come from npm, PyPI, OCI, NuGet or MCPB |
| Namespace ownership | Info | Name verified via GitHub account kanywst |
| Registry text aimed at the AI | Pass | Description and settings contain no hidden characters or instructions to the AI |
| Code red flags | Info | oci packages are not read yet |
Popularity & health: 45 / 100
Worked out from public numbers and our own checks, not from user reviews or ratings. It says how used and how looked-after a server is, not whether it is safe.
0 forks, 59 merged pull requests. Sources: GitHub, npm, PyPI (pypistats.org) and Hacker News (Algolia), fetched 2026-10-31.
What people say
No Hacker News threads link to its source code.
Look further: search Reddit, search Hacker News, read its GitHub issues. These are not reviews we checked.
Facts
| Registry name | io.github.kanywst/mcp-opa-authz |
|---|---|
| Publisher | GitHub account kanywst |
| How the name is verified | Name verified via GitHub account kanywst. The registry ties the name prefix to that account or domain; it does not check the code. |
| Type | Local: runs on your machine from a package |
| Packages | oci ghcr.io/kanywst/mcp-opa-authz:0.6.2 no version given (stdio) |
| Version | 0.6.2 |
| Published | |
| Last updated | |
| Source repository | https://github.com/kanywst/mcp-opa-authz |
| Website | https://github.com/kanywst/mcp-opa-authz#readme |
| Category | Agent safety & governance (assigned by keyword rule, not by hand) |
Related servers in Agent safety & governance
- delego7/7Intent-bound action authorization for AI agents: policy, human…
- Vantio Gate7/7Vantio Gate — Policy Latch dry-run evaluate for agents. No live…
Vulnify8/8Ask Vulnify before an AI agent acts. Check, record, and read policy…
- PriorSeal4/4Bounded authorization and independently verifiable execution…
- atlas-mcp7/7Pre-action authorization for AI agents: permit, deny or escalate…
- Living Stack Community5/6Free seven-tool proof loop for scoped authorization, outcome…