Question: would signed MCP tool-call receipts be useful here? #12140
Replies: 2 comments
|
Hi @cullenmeyers your project looks interesting, currently Unleash has audit logs and change requests in the enterprise version to protect from agents, also all calls from our MCP are authorized by a specific service account token. Nonetheless your project seems interesting but we don't have time at the moment to look into this. |
0 replies
|
Hi Gastón,
Thanks — that makes sense.
If Unleash already has enterprise audit logs, change requests, and
service-account authorization around MCP calls, then AgentReceipt may not
add much unless there is a need for a portable receipt that can be verified
outside the Unleash system.
That helps clarify the boundary for me: AgentReceipt should not duplicate
internal audit logs; it only becomes useful when the record needs to cross
a trust boundary.
Appreciate you taking a look.
…On Mon, Jun 1, 2026 at 9:50 PM Gastón Fournier ***@***.***> wrote:
Hi @cullenmeyers <https://github.com/cullenmeyers> your project looks
interesting, currently Unleash has audit logs and change requests in the
enterprise version to protect from agents, also all calls from our MCP are
authorized by a specific service account token. Nonetheless your project
seems interesting but we don't have time at the moment to look into this.
—
Reply to this email directly, view it on GitHub
<#12140?email_source=notifications&email_token=B3KDSV24TZPW4HMZJXDB22T452BM7A5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCNZRGQ4DCOBZUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSWGM33PORSXEX3DNRUWG2Y#discussioncomment-17148189>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/B3KDSV3LUTWG2KVYYUWPEUT452BM7AVCNFSM6AAAAACZTDL7AKVHI2DSMVQWIX3LMV43URDJONRXK43TNFXW4Q3PNVWWK3TUHMYTOMJUHAYTQOI>
.
Triage notifications, keep track of coding agent tasks and review pull
requests on the go with GitHub Mobile for iOS
<https://github.com/notifications/mobile/ios/B3KDSVZOIHA5QZUBMBU3CCD452BM7A5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCNZRGQ4DCOBZUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSVGM33PORSXEX3JN5ZQ>
and Android
<https://github.com/notifications/mobile/android/B3KDSV3YDDUOP6SNVZB5T23452BM7A5CNFSNUABIM5UWIORPF5TWS5BNNB2WEL2ENFZWG5LTONUW63SDN5WW2ZLOOQXTCNZRGQ4DCOBZUZZGKYLTN5XKO3LFNZ2GS33OUVSXMZLOOSXGM33PORSXEX3BNZSHE33JMQ>.
Download it today!
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
Hi — I found this MCP server and wanted to ask a quick architecture/product question.
I’m building a small open-source project called AgentReceipt for structured, verifiable action records around AI agent/MCP tool calls.
The idea is not to replace normal logs or observability. It is more for trust-boundary actions where someone may later need to prove what happened: caller X invoked tool Y with args Z, the server/tool boundary attested the result, and the receipt can be verified later.
This repo seemed relevant because feature flags / rollout configuration can affect production behavior, so MCP tool calls around flag or environment changes may need more than ordinary debug logs.
Repo for context: https://github.com/cullenmeyers/AgentReceipt
Would this kind of server-side receipt/provenance layer be useful around Unleash MCP tool calls, or would existing Unleash audit logs / normal observability already cover the need?
No worries if this is out of scope — I’m mainly trying to learn where this fits or does not fit.
All reactions