Hi Yu (and team),

Thanks for following up. To clarify my previous message it was
misunderstood : I haven't been idle or working on something separate , I’ve
been fully focused on this same project! Following guidance from Larry and
working with the team, I took time to review our architecture, implement a
whole new design, and rewrite the codebase.

While we have active contributors actively reviewing the architecture and
code, we really need a *Mentor-level review* at this stage to confirm we
are on the right path before moving further.

Here is a summary of where things stand and the key links for review:
Current Project Status & Aegis 2.0 Overview

   -

   *Architecture:* Rebuilt Aegis around a declarative MCP proxy with a thin
   Knox extension (deployed through /ext) rather than per-service adapters.
   -

   *Review Branch:* The 2.0 rewrite is on the review branch:
   aegis-mcp-gateway-2.0-review-ready
   
<https://github.com/vaquarkhan/aegis-mcp-gateway/tree/aegis-mcp-gateway-2.0-review-ready>
   .
   -

   *Scope & Features:* Includes 41 feature families, 31 documented security
   properties, 25 adapter configuration profiles defining 116 MCP tools, 7
   executable reference configurations, and 28 tutorials/guides.
   -

   *Shared State & Persistence:* Shares approval, rate/budget windows, and
   signing state via PostgreSQL when instances share a namespace.

Links for Review & Discussion

   -

   *GitHub Issue for Discussion:* Issue #34
   <https://github.com/vaquarkhan/aegis-mcp-gateway/issues/34>
   -

   *Google Doc Discussion:* Architecture Discussion Doc
   
<https://docs.google.com/document/d/1V0Qf4J1bbWxhT9bp7Azvgzlv1LexBuWmPYMyIApVQlA/edit?tab=t.0&authuser=2>
   -

   *Documentation Site:* Aegis MCP Gateway Docs
   <https://vaquarkhan.github.io/aegis-mcp-gateway/>

*In short:* The team is actively reviewing the code and discussing the
design, which you can see in the branch comments, new issues, bugs etc . We
are keeping our internal discussions in a separate group rather than the
incubator list.. We are also waiting for guidance and next steps from
Apache. .

  Please take a look at the Google Doc or the GitHub issue and share your
thoughts. Your high-level review and feedback on this direction are greatly
appreciated!

Best regards,

Viquar Khan

On Fri, Oct 23, 2026 at 4:58 AM Yu Xiao <[email protected]> wrote:

> Hi Viquar,
> I’m most interested in learning about this project is its core value
> and whether it will have enough active contributors going forward.
> I’ve been following this project, and as you mentioned, development
> has stalled because you’ve been busy—that’s my biggest concern right
> now.
> Joining the incubator is just the first step.
> Best wishes!
>
> Yu Xiao
> ASF Member
> Apache ShenYu V.P.
>
> vaquar khan <[email protected]> 于2026年10月14日周三 06:11写道:
> >
> > Hi all,
> >
> > Sorry for the short break. I spent the last month on a major
> architectural
> > rewrite of Aegis, and I wanted the code to be ready before replying.
> >
> > Larry, thank you for your guidance. It gave the project its new shape.
> You
> > were right that per-service adapters would not scale and that Knox
> already
> > covered more than I realized. I corrected those assumptions and rebuilt
> > Aegis around a declarative MCP proxy with a thin Knox extension.
> >
> > Thanks also to Greg, Andrew, and Yu Xiao for your questions
> > ,feedback,support and to PJ for active contributions  and help to build
> the
> > community and all active contributors .I have added Git issue for review
> > and discussion -
> https://github.com/vaquarkhan/aegis-mcp-gateway/issues/34
> >
> > *Where things are:*
> >
> >    - main still has the original 0.1 code. The 2.0 work is on the review
> >    branch:
> >
> https://github.com/vaquarkhan/aegis-mcp-gateway/tree/aegis-mcp-gateway-2.0-review-ready
> >    - Knox cannot be reused directly because of its heavy dependencies.
> >    Instead, Aegis now has two front doors on the same core: its own
> host, and
> >    an out-of-tree Knox extension deployed through /ext. It is not a Knox
> fork,
> >    and the extension packaging leaves Knox implementation classes.
> >    - The per-engine adapters became declarative HTTP profiles.
> >
> >
> > *What is in the branch:*
> >
> >    - 41 numbered feature families, plus five lettered subfeatures, and 31
> >    documented security properties. These cover identity, authorization,
> >    approvals, rate and budget controls, audit records, and protection
> against
> >    rollback of signed catalogs and related governance documents.
> >    - 25 adapter configuration profiles defining 116 MCP tools. The
> >    default-build tests load and count all 25 configurations, exercise
> HTTP
> >    tools through governance, and test converter tools separately.
> Live-product
> >    evidence varies by profile.
> >    - Seven executable reference configurations and one teaching template
> >    for creating another configuration.
> >    - Historical evidence from 28 earlier runs through both the Aegis own
> >    host and its Knox extension. This includes 27 adopter profiles and one
> >    Spark and Iceberg scenario. These runs were not repeated against the
> >    current review candidate.
> >    - Atlas, Cassandra, Impala, and Paimon have contract-tested
> >    configurations, but no retained live-product runs.
> >    - A step-by-step Hello World guide, 27 project tutorials, and a
> separate
> >    Spark and Iceberg tutorial.
> >    - Built-in token exchange works only for JWT-authenticated MCP
> requests
> >    to the Aegis own host. It fails closed through Knox and for callers
> without
> >    a verified JWT.
> >    - When Aegis instances use the same PostgreSQL database and namespace,
> >    they can share pending approvals, single-use approval and receipt
> state,
> >    and rate and budget windows. Using approvals or receipts across
> instances
> >    also requires matching gateway identity and signing settings.
> PostgreSQL
> >    stores rollback high-water serials for signed server catalogs, catalog
> >    manifests, and break-glass orders.
> >
> >
> > Docs: https://vaquarkhan.github.io/aegis-mcp-gateway/
> >
> > I would appreciate any feedback.
> >
> > Thanks,
> > Viquar Khan
> > https://www.linkedin.com/in/vaquar-khan-b695577
> >
> >
> > On Fri, Sep 18, 2026 at 12:53 AM vaquar khan <[email protected]>
> wrote:
> >
> > > Hi Larry,
> > >
> > > Thank you, this was the most useful review the note has had, and it
> > > genuinely reshaped it for the better.
> > >
> > > First, you were right on the Knox conclusions. I went back and
> re-checked
> > > every claim against the Knox source instead of reasoning about it, and
> > > three were wrong, all of them cases where I said Knox lacked something
> it
> > > actually has: rate limiting (the webappsec DoSFilter), failover and
> retry
> > > (the HA dispatch), and per-service gating (the AclsAuthz provider).
> Those
> > > are fixed, and I have marked each correction so the history is clear.
> > >
> > > To make this easier to collaborate on than email threads or GitHub
> PRs, I
> > > have moved the corrected note into a Google Doc, please comment inline:
> > >
> > >
> https://docs.google.com/document/d/1V0Qf4J1bbWxhT9bp7Azvgzlv1LexBuWmPYMyIApVQlA/edit?tab=t.0
> > >
> > > *On your main question, what a separate piece buys over extending Knox,
> > > here is the short version, with the full evidence in the doc.*
> > >
> > > The core reason is that the two systems work on different planes. Knox
> > > authorizes the transport envelope: I checked AclsAuthorizationFilter,
> it
> > > decides on the principal, groups, and remote IP, and never reads the
> > > request body. MCP puts the operation, the tool name, and the arguments
> > > inside the JSON-RPC body on a single endpoint, so tools/list and a
> > > destructive tools/call are the same URL and method. Knox's own routing
> > > primitive, chains.match on the URL template, cannot tell them apart.
> > > Governing MCP means reading and deciding on the body of every call,
> keeping
> > > and checking per-session state, and screening content, which is a
> different
> > > job from routing REST by URL.
> > >
> > > The threat model is also inverted. Knox authenticates a caller and then
> > > trusts it for the route. Aegis has to assume the caller; the model may
> be
> > > manipulated by prompt injection and keep distrusting the content of
> every
> > > tool call, which is why per-request approval and injection screening
> exist.
> > > "Authenticate then trust" is hard to safely retrofit into "authenticate
> > > then keep distrusting each call"; that is a foundational assumption,
> not a
> > > filter you add on top.
> > >
> > > And to be clear, this is not Aegis instead of Knox. Where our needs
> > > overlap I want to consume Knox, not fork it. KnoxIDF is exactly the
> > > identity foundation I want to build on, federated OIDC plus RFC 8693
> token
> > > exchange with nested act claims is precisely what replaces the token
> > > pass-through Aegis does today, so I would love to align on that and on
> the
> > > AAuth and ID-JAG/XAA/EMA direction. The shape I am proposing is a
> portable
> > > governance core with a thin Knox /ext adapter, so Knox operators get
> this
> > > in-process with your deployment base, while teams not on Knox can
> still run
> > > it. That gives you the integration you are pointing at without folding
> a
> > > distinct, fast-moving governance layer into the gateway's release
> train.
> > >
> > > I have kept the doc honest about what the facts do and do not settle:
> they
> > > establish that this governance layer is a distinct data plane, I have
> > > proposed a protocol-fit prototype (mapping the MCP surface onto Knox's
> > > dispatch model) so we decide that on evidence rather than opinion.
> > >
> > > Thanks again for the guidance and for mentoring this. It has made the
> > > direction much sharper, whichever way we land.
> > >
> > > Best regards,
> > > Viquar Khan
> > >
> > >
> > > On Thu, Sep 17, 2026 at 8:27 AM larry mccay <[email protected]> wrote:
> > >
> > >> Hi Viquar -
> > >>
> > >> The note looks good at a high level.
> > >> I wish github allowed comments on things like this without having to
> > >> create
> > >> a PR. :)
> > >> There are a bunch of Knox comparison conclusions that are not actually
> > >> true.
> > >> Most of what is claimed as not supported is actually supported.
> > >>
> > >> The general strategic goals are well defined though.
> > >>
> > >> I'm curious exactly what additional value having a separate server
> from
> > >> Knox brings.
> > >> Maybe we could just use the same server with this as a separate module
> > >> that
> > >> can be added to the /ext directory to extend the server.
> > >> This would buy us a large existing deployment base with all of the
> > >> provider
> > >> and gateway services built in right out of the box.
> > >>
> > >> In addition, maybe I mentioned this before, Knox 3.1.0 will have a new
> > >> service called KnoxIDF that allows for Knox to be a federated OIDC
> > >> provider
> > >> as well as targeting agentic identity related things.
> > >> The federation will allow for native integration with enterprise IDPs
> from
> > >> keycloak to Ping, to Okta and Entra.
> > >> Delegated Authentication provides RFC 8693 token exchange support (as
> your
> > >> note recognizes) with nested act claims for tracking the chain of
> actors
> > >> on
> > >> behalf of some identity.
> > >> There will be closer alignment with emerging standards like AAuth and
> > >> ID-JAG/XAA/EMA for agentic authorization and governance.
> > >>
> > >> Anyway, just a thought.
> > >>
> > >> You are on the right track either way!
> > >>
> > >> thanks,
> > >>
> > >> --larry
> > >>
> > >>
> > >> On Thu, Sep 17, 2026 at 12:41 AM vaquar khan <[email protected]>
> > >> wrote:
> > >>
> > >> > Hello everyone,
> > >> >
> > >> > Quick follow-up on my previous note  both items are now complete:
> > >> >
> > >> > ICLA field. I have signed and sent the Individual Contributor
> License
> > >> > Agreement to the ASF Secretary.
> > >> >
> > >> > Knox mapping note published. I have written up the short document
> > >> mapping
> > >> > the Aegis governance steps onto the Apache Knox provider chain. It
> > >> > describes how the two share the same shape (entry point → ordered
> policy
> > >> > chain → backend), where Aegis can reuse Knox concepts, and a phased
> > >> plan.
> > >> > It is in the repo here:
> > >> >
> > >> >
> > >>
> https://github.com/vaquarkhan/aegis-mcp-gateway/blob/main/docs/NOTE-aegis-and-apache-knox.md
> > >> >
> > >> > I would welcome your review of the note whenever you have a chance,
> and
> > >> I
> > >> > am happy to adjust it based on community feedback.
> > >> >
> > >> > Best regards,
> > >> > Viquar Khan
> > >> >
> > >> > On Wed, Sep 9, 2026 at 3:55 AM Yu Xiao <[email protected]> wrote:
> > >> >
> > >> > > Hi,
> > >> > >
> > >> > > I'm somewhat interested in this project. It does something
> similar to
> > >> > > the MCP plugin for Apache ShenYu.
> > >> > > If you're looking for a mentor, feel free to add me.
> > >> > >
> > >> > > Best wishes!
> > >> > >
> > >> > > Yu Xiao
> > >> > > ASF Member
> > >> > > Apache ShenYu V.P.
> > >> > >
> > >> > > vaquar khan <[email protected]> 于2026年9月9日周三 14:52写道:
> > >> > > >
> > >> > > > Hi Greg,
> > >> > > >
> > >> > > > Thank you for formatting and uploading the proposal to the
> cwiki; it
> > >> > > looks
> > >> > > > great!
> > >> > > >
> > >> > > > I want to acknowledge your message and confirm that I am
> currently
> > >> > > working
> > >> > > > on the next steps:
> > >> > > >
> > >> > > >  - I have reviewed the published cwiki proposal and it looks
> > >> awesome.
> > >> > > >  - I will file the Individual Contributor License Agreement
> (ICLA)
> > >> as
> > >> > you
> > >> > > > suggested.
> > >> > > >  - I am also working on a short document mapping the Aegis
> > >> governance
> > >> > > steps
> > >> > > > onto the Knox provider chain will update in github repo
> documents
> > >> for
> > >> > > > review.
> > >> > > >
> > >> > > > I will follow up once these are completed. Thanks again for your
> > >> > > guidance!
> > >> > > >
> > >> > > > Best regards,
> > >> > > > Viquar Khan
> > >> > > >
> > >> > > > On Tue, Sep 8, 2026 at 8:43 PM Greg Stein <[email protected]>
> wrote:
> > >> > > >
> > >> > > > > Oooh. It came out nicely. Thank you, Grok.
> > >> > > > >
> > >> > > > >
> > >> > > > >
> > >> > >
> > >> >
> > >>
> https://cwiki.apache.org/confluence/spaces/INCUBATOR/pages/451974017/AegisProposal
> > >> > > > >
> > >> > > > >
> > >> > > > > On Tue, Sep 8, 2026 at 8:21 PM Greg Stein <[email protected]>
> > >> wrote:
> > >> > > > >
> > >> > > > > > Hey Viquar,
> > >> > > > > >
> > >> > > > > > Apologies for my absence. I've been focused on some
> RAI/LLMAO
> > >> work
> > >> > > over
> > >> > > > > > the past few weeks, getting that project launched.
> > >> > > > > >
> > >> > > > > > Let's get your proposal into cwiki. I'll take your original
> > >> > proposal
> > >> > > and
> > >> > > > > > get it formatted and placed in cwiki. I saw that you've got
> > >> further
> > >> > > > > updates
> > >> > > > > > after the discussion with Larry. We can figure that out as
> Step
> > >> > Two.
> > >> > > > > > Regardless of this process outcome, it would be helpful if
> you
> > >> > could
> > >> > > file
> > >> > > > > > an ICLA [1]. An easy step to grease any future wheels.
> > >> > > > > >
> > >> > > > > > I'll reply once I have a cwiki copy of your proposal.
> > >> > > > > >
> > >> > > > > > Cheers,
> > >> > > > > > -g
> > >> > > > > >
> > >> > > > > > [1]
> https://www.apache.org/licenses/contributor-agreements.html
> > >> > > > > >
> > >> > > > > >
> > >> > > > > > On Thu, Aug 27, 2026 at 2:49 PM vaquar khan <
> > >> > [email protected]>
> > >> > > > > > wrote:
> > >> > > > > >
> > >> > > > > >> Hi everyone,
> > >> > > > > >>
> > >> > > > > >> Just checking in on this thread to see where we stand. I
> want
> > >> to
> > >> > > make
> > >> > > > > sure
> > >> > > > > >> I am not holding anything up.
> > >> > > > > >> Greg and Larry, is there anything pending from my side
> that you
> > >> > > need me
> > >> > > > > to
> > >> > > > > >> tackle? Just let me know what the next steps are or if you
> just
> > >> > > need me
> > >> > > > > to
> > >> > > > > >> sit tight for now.
> > >> > > > > >>
> > >> > > > > >> Thanks,
> > >> > > > > >> Viquar Khan
> > >> > > > > >>
> > >> > > > > >> On Mon, Aug 17, 2026 at 11:37 AM vaquar khan <
> > >> > > [email protected]>
> > >> > > > > >> wrote:
> > >> > > > > >>
> > >> > > > > >> > Hi everyone,
> > >> > > > > >> >
> > >> > > > > >> > I want to thank you all again for the invaluable
> feedback and
> > >> > > guidance
> > >> > > > > >> > over the past couple of weeks. It sounds like we have a
> solid
> > >> > > > > consensus
> > >> > > > > >> on
> > >> > > > > >> > the technical direction ,specifically the shift towards a
> > >> > > declarative
> > >> > > > > >> > server definition model for governance context, as Larry
> > >> > > suggested(
> > >> > > > > >> Mentor)
> > >> > > > > >> > . I've also taken note of the feedback regarding growing
> the
> > >> > > committer
> > >> > > > > >> > base, which remains an active priority.
> > >> > > > > >> >
> > >> > > > > >> > With Greg kindly offering to Champion the effort and the
> > >> design
> > >> > > path
> > >> > > > > >> > looking much clearer, I wanted to check in with the wider
> > >> group.
> > >> > > Is
> > >> > > > > >> there
> > >> > > > > >> > anything else anyone would like to discuss, review, or
> > >> clarify
> > >> > > before
> > >> > > > > we
> > >> > > > > >> > officially move to the next stage.
> > >> > > > > >> >
> > >> > > > > >> > Thanks again for all the support and insight.
> > >> > > > > >> >
> > >> > > > > >> > Best regards,
> > >> > > > > >> > Viquar Khan
> > >> > > > > >> >
> > >> > > > > >> > On Thu, Aug 13, 2026 at 3:19 PM vaquar khan <
> > >> > > [email protected]>
> > >> > > > > >> > wrote:
> > >> > > > > >> >
> > >> > > > > >> >> Hi Larry,
> > >> > > > > >> >> That works for me, and it is the cleaner idea. A
> declarative
> > >> > > server
> > >> > > > > >> >> definition that carries the governance context,
> following
> > >> the
> > >> > > Knox
> > >> > > > > path
> > >> > > > > >> >> from code to declarative, is exactly the shape I want,
> and I
> > >> > will
> > >> > > > > >> write it
> > >> > > > > >> >> up that way for the dev list so we can refine it
> together in
> > >> > the
> > >> > > > > >> project.
> > >> > > > > >> >> On next steps, I would value your read on timing. The
> > >> project
> > >> > is
> > >> > > no
> > >> > > > > >> >> longer a single author. A few people have merged code,
> and
> > >> > there
> > >> > > is
> > >> > > > > >> steady
> > >> > > > > >> >> activity, and I will keep growing that. Do you think we
> are
> > >> at
> > >> > a
> > >> > > > > >> reasonable
> > >> > > > > >> >> point to move toward a vote or would you rather we wait
> a
> > >> > little.
> > >> > > > > >> >> My honest lean is to start sooner rather than later.
> Right
> > >> now,
> > >> > > MCP
> > >> > > > > >> >> governance is still greenfield, so we have a rare
> chance to
> > >> > > design
> > >> > > > > the
> > >> > > > > >> >> protocol usage and the governance together from the
> start.
> > >> If
> > >> > we
> > >> > > come
> > >> > > > > >> late,
> > >> > > > > >> >> end-to-end governance stops being a design choice and
> turns
> > >> > into
> > >> > > a
> > >> > > > > >> large
> > >> > > > > >> >> alignment problem across many servers and teams that
> have
> > >> > already
> > >> > > > > >> shipped,
> > >> > > > > >> >> which is far harder to land.
> > >> > > > > >> >> I completely defer to you and Greg on when and how to
> bring
> > >> > this
> > >> > > > > >> forward.
> > >> > > > > >> >> My only ask is that we keep an eye on the window
> together,
> > >> so
> > >> > the
> > >> > > > > >> project
> > >> > > > > >> >> stays ahead of the curve rather than chasing it.
> > >> > > > > >> >>
> > >> > > > > >> >> Thank you,
> > >> > > > > >> >> Viquar Khan
> > >> > > > > >> >>
> > >> > > > > >> >> On Thu, Aug 13, 2026, 12:40 PM larry mccay <
> > >> [email protected]>
> > >> > > > > wrote:
> > >> > > > > >> >>
> > >> > > > > >> >>> Thank you for the detailed response.
> > >> > > > > >> >>>
> > >> > > > > >> >>> I'm thinking that the governance context could be
> captured
> > >> in
> > >> > > more
> > >> > > > > of
> > >> > > > > >> a
> > >> > > > > >> >>> declarative mcp server definition rather than requiring
> > >> > bespoke
> > >> > > code
> > >> > > > > >> for
> > >> > > > > >> >>> each one.
> > >> > > > > >> >>> This makes for a very flexible and more easily
> maintainable
> > >> > > > > >> >>> implementation.
> > >> > > > > >> >>> Knox went through a similar migration from requiring
> code
> > >> to
> > >> > > add a
> > >> > > > > new
> > >> > > > > >> >>> service to dispatch to - to a declarative service
> > >> definition
> > >> > > model.
> > >> > > > > >> >>>
> > >> > > > > >> >>> At any rate, I think that we can move forward based on
> your
> > >> > > response
> > >> > > > > >> and
> > >> > > > > >> >>> your willingness to revisit some of that.
> > >> > > > > >> >>>
> > >> > > > > >> >>> Thanks again,
> > >> > > > > >> >>>
> > >> > > > > >> >>> --larry
> > >> > > > > >> >>>
> > >> > > > > >> >>> On Thu, Aug 13, 2026 at 12:38 PM vaquar khan <
> > >> > > > > [email protected]>
> > >> > > > > >> >>> wrote:
> > >> > > > > >> >>>
> > >> > > > > >> >>> > Hi Larry,
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > Thank you for reading the code and laying this out so
> > >> > > plainly. You
> > >> > > > > >> >>> asked
> > >> > > > > >> >>> > for my honest thoughts, so here they are.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > You are right about the parts that do not scale, and
> I
> > >> want
> > >> > > to say
> > >> > > > > >> that
> > >> > > > > >> >>> > clearly before I push back on anything. A hand
> written
> > >> > > adapter for
> > >> > > > > >> >>> every
> > >> > > > > >> >>> > engine is a treadmill, keeping up with each
> project's API
> > >> > > changes
> > >> > > > > is
> > >> > > > > >> >>> real
> > >> > > > > >> >>> > work I already feel, and when an engine ships its
> own MCP
> > >> > > server
> > >> > > > > my
> > >> > > > > >> >>> adapter
> > >> > > > > >> >>> > for it turns into something to track and duplicate.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > Please note - No Apache project as of now officially
> > >> > released
> > >> > > MCP
> > >> > > > > >> >>> server ,
> > >> > > > > >> >>> > correct me if I am wrong.so we have opportunity to
> fix
> > >> this
> > >> > > area
> > >> > > > > but
> > >> > > > > >> >>> agreed
> > >> > > > > >> >>> > if the project wants its own MCP server we should not
> > >> > create a
> > >> > > > > >> blocker.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > On identity and high availability I am not going to
> > >> reinvent
> > >> > > what
> > >> > > > > >> Knox
> > >> > > > > >> >>> > already does well. Your token exchange with the actor
> > >> chain,
> > >> > > and
> > >> > > > > the
> > >> > > > > >> >>> > externalized token state, give a stronger
> attribution and
> > >> > > defense
> > >> > > > > >> story
> > >> > > > > >> >>> > than a single forwarded credential can, because they
> > >> carry
> > >> > the
> > >> > > > > full
> > >> > > > > >> >>> record
> > >> > > > > >> >>> > of who is acting for whom all the way through the
> call.
> > >> That
> > >> > > is
> > >> > > > > >> >>> exactly the
> > >> > > > > >> >>> > direction I want my next phase to take, the 0.2 work
> I
> > >> > already
> > >> > > > > have
> > >> > > > > >> >>> > planned, building on what you have proven rather than
> > >> > starting
> > >> > > > > from
> > >> > > > > >> >>> > scratch. Tool and resource level authorization also
> fits
> > >> the
> > >> > > way
> > >> > > > > the
> > >> > > > > >> >>> > gateway already makes its decisions, so aligning it
> with
> > >> the
> > >> > > Knox
> > >> > > > > >> >>> model is
> > >> > > > > >> >>> > a natural step rather than a rebuild, and a discovery
> > >> > catalog
> > >> > > sits
> > >> > > > > >> >>> > comfortably on top of that. Those are easy yeses.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > Where I see it differently is the idea that proxying
> MCP
> > >> > > servers
> > >> > > > > can
> > >> > > > > >> >>> be the
> > >> > > > > >> >>> > whole story. I have been thinking hard about this
> and I
> > >> do
> > >> > not
> > >> > > > > think
> > >> > > > > >> >>> it can
> > >> > > > > >> >>> > be, for a reason that is central to why the project
> > >> exists.
> > >> > > The
> > >> > > > > >> moment
> > >> > > > > >> >>> the
> > >> > > > > >> >>> > gateway only forwards calls to an opaque server, it
> can
> > >> no
> > >> > > longer
> > >> > > > > >> >>> reason
> > >> > > > > >> >>> > about what those calls do, and most of the value I
> care
> > >> > about
> > >> > > > > >> depends
> > >> > > > > >> >>> on
> > >> > > > > >> >>> > exactly that reasoning. Deciding that a call is read
> > >> only,
> > >> > or
> > >> > > > > that a
> > >> > > > > >> >>> result
> > >> > > > > >> >>> > needs redacting, or that an operation is destructive
> and
> > >> > needs
> > >> > > > > >> >>> approval,
> > >> > > > > >> >>> > all need some understanding of the tool, not just its
> > >> > bytes. A
> > >> > > > > pure
> > >> > > > > >> >>> proxy
> > >> > > > > >> >>> > governs the envelope, not the action.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > MCP does give us a little to work with, because a
> tool
> > >> can
> > >> > > > > advertise
> > >> > > > > >> >>> hints
> > >> > > > > >> >>> > about whether it is read only or destructive. But
> those
> > >> > hints
> > >> > > come
> > >> > > > > >> >>> from the
> > >> > > > > >> >>> > very server we are trying to govern, so the safe
> design
> > >> is
> > >> > to
> > >> > > > > treat
> > >> > > > > >> >>> them as
> > >> > > > > >> >>> > claims to verify, not facts to trust. Otherwise a
> > >> careless
> > >> > or
> > >> > > > > >> hostile
> > >> > > > > >> >>> > server marks a destructive tool as harmless and walks
> > >> > straight
> > >> > > > > >> through.
> > >> > > > > >> >>> > That is a governance decision a forwarding proxy
> cannot
> > >> make
> > >> > > on
> > >> > > > > its
> > >> > > > > >> >>> own,
> > >> > > > > >> >>> > and I think it is the kind of thing our layer has to
> own.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > This is not hypothetical. Tool poisoning is near the
> top
> > >> of
> > >> > > the
> > >> > > > > >> OWASP
> > >> > > > > >> >>> MCP
> > >> > > > > >> >>> > Top 10, and it works by hiding the payload in the
> tool
> > >> > > description
> > >> > > > > >> or
> > >> > > > > >> >>> > schema that the model reads as trusted instructions,
> with
> > >> > rug
> > >> > > > > pulls
> > >> > > > > >> >>> doing
> > >> > > > > >> >>> > the same thing at runtime by mutating the tools list
> > >> after
> > >> > the
> > >> > > > > >> server
> > >> > > > > >> >>> was
> > >> > > > > >> >>> > approved. Those live in the protocol responses, so
> they
> > >> slip
> > >> > > past
> > >> > > > > >> >>> static
> > >> > > > > >> >>> > checks entirely, and the only thing that catches
> them is
> > >> a
> > >> > > layer
> > >> > > > > >> that
> > >> > > > > >> >>> pins
> > >> > > > > >> >>> > each tool and rejects it the moment its definition
> > >> changes.
> > >> > I
> > >> > > have
> > >> > > > > >> been
> > >> > > > > >> >>> > close to this problem at the library level for a
> while
> > >> now
> > >> > > with
> > >> > > > > the
> > >> > > > > >> >>> > MCP-Bastion
> > >> > > > > >> >>> > packages <https://github.com/vaquarkhan/MCP-Bastion>
> I
> > >> > > maintain,
> > >> > > > > >> >>> which do
> > >> > > > > >> >>> > injection blocking, supply chain checksums, and
> redaction
> > >> > > around a
> > >> > > > > >> >>> single
> > >> > > > > >> >>> > server, and Aegis is really the governed, multi
> engine
> > >> > > version of
> > >> > > > > >> that
> > >> > > > > >> >>> same
> > >> > > > > >> >>> > thinking. That background is a big part of why I am
> > >> cautious
> > >> > > about
> > >> > > > > >> >>> treating
> > >> > > > > >> >>> > a plain proxy as enough.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > There is a second reason proxying and the identity
> work
> > >> are
> > >> > > really
> > >> > > > > >> one
> > >> > > > > >> >>> > design and not two, and you know this side far better
> > >> than I
> > >> > > do. A
> > >> > > > > >> >>> gateway
> > >> > > > > >> >>> > that discovers a server and forwards a caller's
> token is
> > >> the
> > >> > > > > classic
> > >> > > > > >> >>> > confused deputy, which is why the MCP spec forbids
> token
> > >> > > > > >> passthrough,
> > >> > > > > >> >>> and
> > >> > > > > >> >>> > doing it safely means re-minting identity at the
> > >> boundary,
> > >> > > which
> > >> > > > > is
> > >> > > > > >> >>> exactly
> > >> > > > > >> >>> > your token exchange work. So I do not see proxy
> first and
> > >> > > > > delegated
> > >> > > > > >> >>> > identity as separate tracks. The identity model is
> what
> > >> > makes
> > >> > > > > >> proxying
> > >> > > > > >> >>> safe
> > >> > > > > >> >>> > in the first place, which is one more reason I want
> to
> > >> build
> > >> > > on
> > >> > > > > >> yours
> > >> > > > > >> >>> > rather than alongside it.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > So here is where I land. Proxy first for reach, yes,
> and
> > >> I
> > >> > > will
> > >> > > > > pull
> > >> > > > > >> >>> the 20
> > >> > > > > >> >>> > or 25 or 29 thin adapters out and lead with the
> proxy.
> > >> But I
> > >> > > want
> > >> > > > > to
> > >> > > > > >> >>> keep
> > >> > > > > >> >>> > the four deep adapters as more than a break glass
> > >> fallback.
> > >> > > They
> > >> > > > > are
> > >> > > > > >> >>> the
> > >> > > > > >> >>> > reference for what real governance looks like when
> the
> > >> > gateway
> > >> > > > > >> actually
> > >> > > > > >> >>> > understands the tool, and the proxy path should be
> > >> measured
> > >> > > > > against
> > >> > > > > >> >>> that
> > >> > > > > >> >>> > bar rather than settling below it. I would also keep
> the
> > >> > data
> > >> > > > > >> platform
> > >> > > > > >> >>> as
> > >> > > > > >> >>> > the proving ground, not because governance is data
> > >> specific,
> > >> > > but
> > >> > > > > >> >>> because
> > >> > > > > >> >>> > the hardest cases live there: destructive schema
> changes,
> > >> > bulk
> > >> > > > > data
> > >> > > > > >> >>> leaving
> > >> > > > > >> >>> > the building, personal data showing up in a query
> > >> result. If
> > >> > > the
> > >> > > > > >> design
> > >> > > > > >> >>> > holds against those, it holds anywhere.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > None of this is me digging in. I could be wrong
> about how
> > >> > > much of
> > >> > > > > >> that
> > >> > > > > >> >>> > semantic depth is worth the cost, and that is
> genuinely
> > >> > where
> > >> > > I
> > >> > > > > >> would
> > >> > > > > >> >>> value
> > >> > > > > >> >>> > your judgment most, since you have carried a security
> > >> > gateway
> > >> > > > > >> through
> > >> > > > > >> >>> years
> > >> > > > > >> >>> > of this. My next step, if it is useful, is to write
> up
> > >> how
> > >> > the
> > >> > > > > >> >>> governance
> > >> > > > > >> >>> > steps map onto the Knox provider chain and this proxy
> > >> model,
> > >> > > and
> > >> > > > > put
> > >> > > > > >> >>> it on
> > >> > > > > >> >>> > the dev list so we can pull it apart together.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > Mostly I want to thank you. This is the most useful
> > >> feedback
> > >> > > the
> > >> > > > > >> >>> project
> > >> > > > > >> >>> > has had, and having someone who has actually shipped
> and
> > >> > > hardened
> > >> > > > > a
> > >> > > > > >> >>> gateway
> > >> > > > > >> >>> > is exactly what it needs. Please keep pushing on
> this. I
> > >> > would
> > >> > > > > >> rather
> > >> > > > > >> >>> be
> > >> > > > > >> >>> > corrected now than be polite and wrong.
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > Best regards,
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > Viquar Khan
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > https://www.linkedin.com/in/vaquar-khan-b695577/
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > On Wed, Aug 12, 2026 at 11:08 PM Chandan Singh S M <
> > >> > > > > >> >>> > [email protected]>
> > >> > > > > >> >>> > wrote:
> > >> > > > > >> >>> >
> > >> > > > > >> >>> > > Please remove me from the general email list
> > >> > > > > >> >>> > >
> > >> > > > > >> >>> > > On Thu, Aug 13, 2026, 07:34 larry mccay <
> > >> > [email protected]>
> > >> > > > > >> wrote:
> > >> > > > > >> >>> > >
> > >> > > > > >> >>> > >> Hi Vaquar -
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> I've taken a look at the repo and think you have
> > >> > > identified an
> > >> > > > > >> >>> > interesting
> > >> > > > > >> >>> > >> scope for data services and MCP.
> > >> > > > > >> >>> > >> I would like to discuss the technical direction
> for
> > >> this
> > >> > > work.
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> At this point, it seems that the project is
> > >> essentially
> > >> > > > > creating
> > >> > > > > >> an
> > >> > > > > >> >>> uber
> > >> > > > > >> >>> > >> MCP Server that includes a separate set of tools
> and
> > >> > > resources
> > >> > > > > >> for
> > >> > > > > >> >>> each
> > >> > > > > >> >>> > >> data service built directly into the project.
> > >> > > > > >> >>> > >> There are a couple challenges with this approach:
> > >> > > > > >> >>> > >> 1. Scaling to keep up with any and all changes
> made to
> > >> > the
> > >> > > > > >> service
> > >> > > > > >> >>> APIs
> > >> > > > > >> >>> > >> across the release trains of multiple projects.
> > >> > > > > >> >>> > >> 2. Very real possibility of the data services
> creating
> > >> > > their
> > >> > > > > own
> > >> > > > > >> MCP
> > >> > > > > >> >>> > >> Servers that we would need to compete with or
> > >> duplicate
> > >> > and
> > >> > > > > >> track,
> > >> > > > > >> >>> etc.
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> My recommendation is to concentrate on the pure
> > >> > governance
> > >> > > and
> > >> > > > > >> >>> security
> > >> > > > > >> >>> > >> aspects of using MCP Servers in an enterprise
> setting
> > >> > with
> > >> > > > > things
> > >> > > > > >> >>> like:
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> 1. MCP Server Proxy - discover and integrate MCP
> > >> Servers
> > >> > > via
> > >> > > > > the
> > >> > > > > >> MCP
> > >> > > > > >> >>> > >> protocol itself as the primary target
> > >> > > > > >> >>> > >>     * configuration based integration rather than
> code
> > >> > > > > >> >>> > >> 2. Break glass alternative for creating an MCP
> Server
> > >> for
> > >> > > > > >> services
> > >> > > > > >> >>> where
> > >> > > > > >> >>> > >> one doesn't exist
> > >> > > > > >> >>> > >>     * possible "mashup" MCP Servers for
> interesting
> > >> > > > > combinations
> > >> > > > > >> of
> > >> > > > > >> >>> > other
> > >> > > > > >> >>> > >> MCP Servers
> > >> > > > > >> >>> > >> 3. OAuth Discovery and Identity concerns for
> delegated
> > >> > > agentic
> > >> > > > > >> auth
> > >> > > > > >> >>> and
> > >> > > > > >> >>> > >> proper attribution to address agentic identity
> > >> challenges
> > >> > > with
> > >> > > > > >> MCP
> > >> > > > > >> >>> > Servers
> > >> > > > > >> >>> > >> 4. Agentic Authorization - providing fine grained
> > >> access
> > >> > > at the
> > >> > > > > >> >>> tool and
> > >> > > > > >> >>> > >> resource level rather than the server level for
> > >> > > > > >> federated/delegated
> > >> > > > > >> >>> > >> identities
> > >> > > > > >> >>> > >> 5. MCP Server Catalog for user level discovery API
> > >> and UI
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> Note that I have thought through some of the
> above but
> > >> > not
> > >> > > all
> > >> > > > > of
> > >> > > > > >> >>> the
> > >> > > > > >> >>> > >> above. :)
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> These are traditional gateway features that will
> > >> remove
> > >> > the
> > >> > > > > >> burden
> > >> > > > > >> >>> of
> > >> > > > > >> >>> > >> having to maintain bespoke integrations with each
> > >> service
> > >> > > > > >> >>> independently.
> > >> > > > > >> >>> > >> They also generically add value to any MCP Server
> > >> > ecosystem
> > >> > > > > >> whether
> > >> > > > > >> >>> in
> > >> > > > > >> >>> > the
> > >> > > > > >> >>> > >> data platform or any other.
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> What are your thoughts on this take?
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> thanks,
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> --larry
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> On Sun, Aug 9, 2026 at 10:06 AM larry mccay <
> > >> > > [email protected]
> > >> > > > > >
> > >> > > > > >> >>> wrote:
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >> > > And for anybody at the ASF who *has*
> > >> > > > > >> >>> > >> > > implemented an MCP, or *intends* to implement
> an
> > >> MCP,
> > >> > > to at
> > >> > > > > >> >>> least
> > >> > > > > >> >>> > give
> > >> > > > > >> >>> > >> > this
> > >> > > > > >> >>> > >> > > proposal a once-over.
> > >> > > > > >> >>> > >> >
> > >> > > > > >> >>> > >> > I failed to mention that I have created a couple
> > >> MCPs
> > >> > as
> > >> > > well
> > >> > > > > >> as
> > >> > > > > >> >>> a POC
> > >> > > > > >> >>> > >> of
> > >> > > > > >> >>> > >> > Knox as an MCP Gateway. :)
> > >> > > > > >> >>> > >> >
> > >> > > > > >> >>> > >> > On Sat, Aug 8, 2026 at 10:25 PM Justin Mclean <
> > >> > > > > >> >>> > [email protected]
> > >> > > > > >> >>> > >> >
> > >> > > > > >> >>> > >> > wrote:
> > >> > > > > >> >>> > >> >
> > >> > > > > >> >>> > >> >> Hi,
> > >> > > > > >> >>> > >> >> ,
> > >> > > > > >> >>> > >> >> > I wanted to say a word on project governance.
> > >> > > > > Specifically,
> > >> > > > > >> >>> how can
> > >> > > > > >> >>> > >> we
> > >> > > > > >> >>> > >> >> > ensure that individuals who are already PPMC
> > >> members
> > >> > > > > across
> > >> > > > > >> >>> > multiple
> > >> > > > > >> >>> > >> >> > projects don't monopolize leadership roles
> here?
> > >> > > > > >> >>> > >> >> >
> > >> > > > > >> >>> > >> >> > Are there any existing quotas, guidelines, or
> > >> best
> > >> > > > > practices
> > >> > > > > >> >>> > >> regarding
> > >> > > > > >> >>> > >> >> the
> > >> > > > > >> >>> > >> >> > number of projects someone can be involved
> in as
> > >> a
> > >> > > > > >> committer or
> > >> > > > > >> >>> > PPMC
> > >> > > > > >> >>> > >> >> > member?
> > >> > > > > >> >>> > >> >>
> > >> > > > > >> >>> > >> >> I suggest you take a look at these documents on
> > >> > mentors:
> > >> > > > > >> >>> > >> >> https://cwiki.apache.org/confluence/x/cQqRFw
> > >> > > > > >> >>> > >> >> https://cwiki.apache.org/confluence/x/cwqRFw
> > >> > > > > >> >>> > >> >>
> > >> > > > > >> >>> > >> >> Re (P)PMC or committers, there are no real
> limits
> > >> or
> > >> > > advice
> > >> > > > > >> that
> > >> > > > > >> >>> I’m
> > >> > > > > >> >>> > >> >> aware of.
> > >> > > > > >> >>> > >> >> A few reasons why:
> > >> > > > > >> >>> > >> >>
> > >> > > > > >> >>> > >> >> Committer and PPMC roles are not a fixed pool
> of
> > >> > seats.
> > >> > > > > Adding
> > >> > > > > >> >>> one
> > >> > > > > >> >>> > >> person
> > >> > > > > >> >>> > >> >> doesn't prevent adding another. If several
> > >> experienced
> > >> > > > > >> >>> contributors
> > >> > > > > >> >>> > are
> > >> > > > > >> >>> > >> >> eager to step up, the answer is to invite all
> of
> > >> them,
> > >> > > not
> > >> > > > > to
> > >> > > > > >> >>> ration
> > >> > > > > >> >>> > >> roles
> > >> > > > > >> >>> > >> >> among them.
> > >> > > > > >> >>> > >> >>
> > >> > > > > >> >>> > >> >> The ASF works on merit within each project.
> Merit
> > >> is
> > >> > > earned
> > >> > > > > >> per
> > >> > > > > >> >>> > project
> > >> > > > > >> >>> > >> >> and doesn't carry over, so someone being on
> five
> > >> other
> > >> > > PPMCs
> > >> > > > > >> >>> gives
> > >> > > > > >> >>> > >> them no
> > >> > > > > >> >>> > >> >> standing here beyond what they've contributed.
> > >> > Equally,
> > >> > > it
> > >> > > > > >> >>> doesn't
> > >> > > > > >> >>> > >> count
> > >> > > > > >> >>> > >> >> against them.
> > >> > > > > >> >>> > >> >>
> > >> > > > > >> >>> > >> >> People involved in many projects are usually a
> > >> > benefit,
> > >> > > not
> > >> > > > > a
> > >> > > > > >> >>> > problem.
> > >> > > > > >> >>> > >> >> They bring experience of how other communities
> > >> handle
> > >> > > > > >> releases,
> > >> > > > > >> >>> votes
> > >> > > > > >> >>> > >> and
> > >> > > > > >> >>> > >> >> governance, which is exactly what a new project
> > >> needs.
> > >> > > > > >> >>> > >> >>
> > >> > > > > >> >>> > >> >> If your concern behind your question is
> capacity:
> > >> can
> > >> > > > > someone
> > >> > > > > >> >>> spread
> > >> > > > > >> >>> > >> >> across many projects give this one enough
> > >> attention?
> > >> > > That
> > >> > > > > >> varies
> > >> > > > > >> >>> > >> widely and
> > >> > > > > >> >>> > >> >> on a case-by-case basis.
> > >> > > > > >> >>> > >> >>
> > >> > > > > >> >>> > >> >> Kind Regards,
> > >> > > > > >> >>> > >> >> Justin
> > >> > > > > >> >>> > >> >>
> > >> > > > > >> >>> > >> >>
> > >> > > > > >> >>> > >> >>
> > >> > > > > >> >>> > >>
> > >> > > > > >> >>> > >
> > >> > > > > >> >>> >
> > >> > > > > >> >>>
> > >> > > > > >> >>
> > >> > > > > >>
> > >> > > > > >
> > >> > > > >
> > >> > >
> > >> > >
> ---------------------------------------------------------------------
> > >> > > To unsubscribe, e-mail: [email protected]
> > >> > > For additional commands, e-mail:
> [email protected]
> > >> > >
> > >> > >
> > >> >
> > >>
> > >
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: [email protected]
> For additional commands, e-mail: [email protected]
>
>

Reply via email to