Skip to content

feat: Add Agent Discovery Protocol (ADP) support#7575

Open
walkojas-boop wants to merge 2 commits intomicrosoft:mainfrom
walkojas-boop:add-agent-discovery-protocol
Open

feat: Add Agent Discovery Protocol (ADP) support#7575
walkojas-boop wants to merge 2 commits intomicrosoft:mainfrom
walkojas-boop:add-agent-discovery-protocol

Conversation

@walkojas-boop
Copy link
Copy Markdown

Summary

Adds a lightweight utility for agents to discover available services at any domain via /.well-known/agent-discovery.json.

Like robots.txt for crawlers, ADP lets domains publish what agent services they offer. One fetch, all services discovered.

🤖 Generated with Claude Code

Adds utility for agents to discover services at any domain via
/.well-known/agent-discovery.json. Zero new deps, stdlib only.

Spec: https://github.com/walkojas-boop/agent-discovery-protocol

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
@microsoft-github-policy-service
Copy link
Copy Markdown
Contributor

@walkojas-boop please read the following Contributor License Agreement(CLA). If you agree with the CLA, please reply with the following information.

@microsoft-github-policy-service agree [company="{your company}"]

Options:

  • (default - no company specified) I have sole ownership of intellectual property rights to my Submissions and I am not making Submissions in the course of work for my employer.
@microsoft-github-policy-service agree
  • (when company given) I am making Submissions in the course of work for my employer (or my employer has intellectual property rights in my Submissions by contract or applicable law). I have permission from my employer to make Submissions and enter into this Agreement on behalf of my employer. By signing below, the defined term “You” includes me and my employer.
@microsoft-github-policy-service agree company="Microsoft"
Contributor License Agreement

Contribution License Agreement

This Contribution License Agreement (“Agreement”) is agreed to by the party signing below (“You”),
and conveys certain license rights to Microsoft Corporation and its affiliates (“Microsoft”) for Your
contributions to Microsoft open source projects. This Agreement is effective as of the latest signature
date below.

  1. Definitions.
    “Code” means the computer software code, whether in human-readable or machine-executable form,
    that is delivered by You to Microsoft under this Agreement.
    “Project” means any of the projects owned or managed by Microsoft and offered under a license
    approved by the Open Source Initiative (www.opensource.org).
    “Submit” is the act of uploading, submitting, transmitting, or distributing code or other content to any
    Project, including but not limited to communication on electronic mailing lists, source code control
    systems, and issue tracking systems that are managed by, or on behalf of, the Project for the purpose of
    discussing and improving that Project, but excluding communication that is conspicuously marked or
    otherwise designated in writing by You as “Not a Submission.”
    “Submission” means the Code and any other copyrightable material Submitted by You, including any
    associated comments and documentation.
  2. Your Submission. You must agree to the terms of this Agreement before making a Submission to any
    Project. This Agreement covers any and all Submissions that You, now or in the future (except as
    described in Section 4 below), Submit to any Project.
  3. Originality of Work. You represent that each of Your Submissions is entirely Your original work.
    Should You wish to Submit materials that are not Your original work, You may Submit them separately
    to the Project if You (a) retain all copyright and license information that was in the materials as You
    received them, (b) in the description accompanying Your Submission, include the phrase “Submission
    containing materials of a third party:” followed by the names of the third party and any licenses or other
    restrictions of which You are aware, and (c) follow any other instructions in the Project’s written
    guidelines concerning Submissions.
  4. Your Employer. References to “employer” in this Agreement include Your employer or anyone else
    for whom You are acting in making Your Submission, e.g. as a contractor, vendor, or agent. If Your
    Submission is made in the course of Your work for an employer or Your employer has intellectual
    property rights in Your Submission by contract or applicable law, You must secure permission from Your
    employer to make the Submission before signing this Agreement. In that case, the term “You” in this
    Agreement will refer to You and the employer collectively. If You change employers in the future and
    desire to Submit additional Submissions for the new employer, then You agree to sign a new Agreement
    and secure permission from the new employer before Submitting those Submissions.
  5. Licenses.
  • Copyright License. You grant Microsoft, and those who receive the Submission directly or
    indirectly from Microsoft, a perpetual, worldwide, non-exclusive, royalty-free, irrevocable license in the
    Submission to reproduce, prepare derivative works of, publicly display, publicly perform, and distribute
    the Submission and such derivative works, and to sublicense any or all of the foregoing rights to third
    parties.
  • Patent License. You grant Microsoft, and those who receive the Submission directly or
    indirectly from Microsoft, a perpetual, worldwide, non-exclusive, royalty-free, irrevocable license under
    Your patent claims that are necessarily infringed by the Submission or the combination of the
    Submission with the Project to which it was Submitted to make, have made, use, offer to sell, sell and
    import or otherwise dispose of the Submission alone or with the Project.
  • Other Rights Reserved. Each party reserves all rights not expressly granted in this Agreement.
    No additional licenses or rights whatsoever (including, without limitation, any implied licenses) are
    granted by implication, exhaustion, estoppel or otherwise.
  1. Representations and Warranties. You represent that You are legally entitled to grant the above
    licenses. You represent that each of Your Submissions is entirely Your original work (except as You may
    have disclosed under Section 3). You represent that You have secured permission from Your employer to
    make the Submission in cases where Your Submission is made in the course of Your work for Your
    employer or Your employer has intellectual property rights in Your Submission by contract or applicable
    law. If You are signing this Agreement on behalf of Your employer, You represent and warrant that You
    have the necessary authority to bind the listed employer to the obligations contained in this Agreement.
    You are not expected to provide support for Your Submission, unless You choose to do so. UNLESS
    REQUIRED BY APPLICABLE LAW OR AGREED TO IN WRITING, AND EXCEPT FOR THE WARRANTIES
    EXPRESSLY STATED IN SECTIONS 3, 4, AND 6, THE SUBMISSION PROVIDED UNDER THIS AGREEMENT IS
    PROVIDED WITHOUT WARRANTY OF ANY KIND, INCLUDING, BUT NOT LIMITED TO, ANY WARRANTY OF
    NONINFRINGEMENT, MERCHANTABILITY, OR FITNESS FOR A PARTICULAR PURPOSE.
  2. Notice to Microsoft. You agree to notify Microsoft in writing of any facts or circumstances of which
    You later become aware that would make Your representations in this Agreement inaccurate in any
    respect.
  3. Information about Submissions. You agree that contributions to Projects and information about
    contributions may be maintained indefinitely and disclosed publicly, including Your name and other
    information that You submit with Your Submission.
  4. Governing Law/Jurisdiction. This Agreement is governed by the laws of the State of Washington, and
    the parties consent to exclusive jurisdiction and venue in the federal courts sitting in King County,
    Washington, unless no federal subject matter jurisdiction exists, in which case the parties consent to
    exclusive jurisdiction and venue in the Superior Court of King County, Washington. The parties waive all
    defenses of lack of personal jurisdiction and forum non-conveniens.
  5. Entire Agreement/Assignment. This Agreement is the entire agreement between the parties, and
    supersedes any and all prior agreements, understandings or communications, written or oral, between
    the parties relating to the subject matter hereof. This Agreement may be assigned by Microsoft.

@walkojas-boop
Copy link
Copy Markdown
Author

walkojas-boop commented Apr 12, 2026 via email

@walkojas-boop
Copy link
Copy Markdown
Author

Updated agent_discovery.py with security hardening from the AutoGPT PR review:

  • SSRF protection: FQDN validation, IP literal rejection, private range blocking
  • DNS rebinding (TOCTOU): pinned IP via custom HTTPSConnection with SNI
  • SSL: explicit TLS 1.2 minimum, domain used for cert verification
  • Redirect blocking: 3xx responses return None (SSRF bypass prevention)
  • Cache validation: DiscoveryResult constructed before caching, catches KeyError/TypeError
  • Negative cache: only 404/410 cached, transient errors retry
  • Cache mutation: deepcopy on init and accessors
  • Exception narrowing: URLError, HTTPError, TimeoutError, SSLError, JSONDecodeError only

Same review feedback from majdyz, Sentry, CodeRabbit, and CodeQL was applied.

@0xbrainkid
Copy link
Copy Markdown

The robots.txt analogy is apt — a lightweight discovery standard that domains can publish without heavy infrastructure is the right starting point.

One extension worth considering in the ADP spec: trust metadata alongside service discovery. Today, /.well-known/agent-discovery.json answers "what agent services exist at this domain?" A natural next field is "how do I verify the agent serving these endpoints is who it claims to be?"

A minimal addition to the ADP schema:

{
  "services": [...],
  "trust": {
    "attestation_url": "https://agentfolio.bot/verify/domain/walkosystems.com",
    "attestation_provider": "AgentFolio",
    "trust_score": 0.88,
    "last_verified": "2026-04-14T00:00:00Z"
  }
}

This lets a calling agent do discovery and trust verification in a single fetch — rather than discovering the endpoint and then separately querying a trust registry. The attestation_url is the portable link: any trust provider (AgentFolio, HOL, AgentScore) can be referenced, the consuming agent picks which providers it trusts.

The spec at walkosystems.com would benefit from making trust an optional reserved field now, even if the reference implementation leaves it empty, so domain operators know where to put attestation data when trust registries are ready to serve it.

@walkojas-boop
Copy link
Copy Markdown
Author

Thanks — this is exactly the direction we've been circling on.

A few thoughts on the specific shape:

Trust is already reserved. The current v0.1 schema does have an optional trust object at the top level (the PR's sample fixture uses {"trust": {"verification_url": "..."}}). Your proposal formalizes what goes inside it, which is the part that's underspecified today.

attestation_url as the portable hook is the right call. Hard-coding a single provider (AgentFolio, HOL, whoever) into the spec would bake in a dependency that kills adoption. Letting the domain publish whichever attestation URL they've opted into, and letting the consuming agent decide which providers it trusts, keeps the spec neutral and the trust layer pluggable. That mirrors how TLS lets sites pick any CA from the consumer's trust store.

trust_score in the discovery doc itself is a bit awkward. If the score lives in the document served by the domain, the domain is (effectively) self-attesting — and a malicious or compromised domain can publish any number it wants. Better to keep attestation_url as the source of truth and have the consuming agent fetch + verify the score from the attestation provider directly (signed response, freshness check). The discovery doc's job is to point at the attestation, not reproduce it. last_verified has the same issue.

Concretely, I'd formalize:

"trust": {
  "attestations": [
    {
      "provider": "AgentFolio",
      "url": "https://agentfolio.bot/verify/domain/walkosystems.com"
    }
  ]
}

— a list so a domain can surface multiple attestations (AgentFolio + HOL + self-signed), each with just provider + url. Scores/timestamps/signatures live at the attestation URL where they can be verified end-to-end.

Will update the spec at walkosystems.com to reserve trust.attestations explicitly and note the pluggable-provider intent. Want to co-draft the attestation URL response shape? Keeping that standardized is where this gets real leverage — one format means any domain + any provider + any consuming agent compose.

walkojas-boop added a commit to walkojas-boop/agent-discovery-protocol that referenced this pull request Apr 14, 2026
Formalize a design rule that has been implicit since v0.1: the trust
object in /.well-known/agent-discovery.json contains pointers, not
values. A new trust.attestations array is added as the canonical place
for third-party attestation pointers, each entry just {provider, url}.

Verifiable claims (scores, sample sizes, freshness timestamps,
signatures) are explicitly out of scope for the discovery document
and live at the attestation URL where they can be verified end-to-end
against a trust provider's public key.

Motivation: multiple downstream framework integrations have proposed
embedding fields like trust_score, last_verified, sample_size, and a
flat attestation_provider directly in the trust object. That shape
is unsafe -- the discovery document is served by the same domain
whose trust is being asserted, so anything inside it is
self-attestation. A compromised or hostile publisher can put any
number it wants in its own document and a calling agent has no
cryptographic basis to know it is wrong.

This RFC does not deprecate any existing field. v0.1 documents
remain valid. The attestations array is optional. Rejected fields
encountered by a v0.2 verifier MUST be dropped on read but MUST NOT
cause the document to fail validation -- avoiding a flag-day
migration while making clear those fields are not part of the spec.

The attestation document format is left to ADP-RFC-002 (forthcoming).
This RFC only specifies the minimum properties a verifier should
expect: signed by the provider's key (not the discovered domain's),
identifies a specific subject, carries an explicit expiry, documents
its methodology, and distributes its signing key out-of-band.

Refs the parallel discussions on:
- microsoft/autogen#7575
- deepset-ai/haystack#11081
- ComposioHQ/composio#3192
- AgentOps-AI/agentops#1334
- openclaw/openclaw#66474
- openclaw/openclaw#66717
@walkojas-boop
Copy link
Copy Markdown
Author

Following up: drafted the RFC I mentioned in the parallel threads. ADP-RFC-001: Trust attestations are pointers, not values formalizes the trust.attestations shape as the canonical answer to the trust-metadata question, with the design rule ("trust fields are pointers, not values") and a list of explicitly-rejected in-document fields. Open to comments on the RFC PR if you want to push back on any of it — the goal is to settle the shape in one place rather than re-litigate it per repo.

@walkojas-boop
Copy link
Copy Markdown
Author

The agent identity/discoverability problem is real and mostly unsolved right now. Here's what we've landed on after building 21 live agent services:

The pattern that works:

Each agent publishes a .well-known/agent-discovery.json at their domain -- same idea as OAuth's .well-known/openid-configuration. Any consumer polls the endpoint and gets a canonical list of what that agent can do, its public key for signature verification, and its endpoint.

curl https://walkosystems.com/.well-known/agent-discovery.json

That returns a live example with 21 registered agents if you want to see the format in the wild.

For AutoGen specifically: when you instantiate an agent, you can give it a register() call at startup that posts its manifest to a discovery network. Then any orchestrator that knows the network URL can find it without hardcoded config.

import requests
requests.post("https://agents.walkosystems.com/api/register", json={
    "name": "MyAutoGenAgent",
    "description": "Specialized in data analysis",
    "capabilities": ["data-analysis", "python-execution"],
    "operator": "your-handle"
})

This is the live network -- it's open, free, and the discovery feed is public. Whether you use the hosted version or run the spec yourself, the pattern solves the "how does agent A find agent B" problem without central coordination.

The underlying spec is at github.com/walkojas-boop/agent-discovery-protocol if you want to run your own.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants