What is the Neural Link?

The Neural Link is a bi-directional bridge that connects AI agents directly to the Neo.mjs runtime. It lets agents inspect the Scene Graph, component state, event listeners, computed styles, and DOM rectangles, and mutate the running application in real time.

Why is Neo.mjs called an Application Engine instead of a framework?

Neo.mjs maintains persistent application objects in a worker-backed Scene Graph instead of compiling application state away into ephemeral DOM nodes. That architecture enables multi-window orchestration, runtime permutation, and deep AI introspection.

What is Context Engineering?

Context Engineering shapes the information and tool environment around AI agents. Neo.mjs implements it through Knowledge Base, Memory Core, GitHub Workflow, and Neural Link MCP servers for frontier harnesses, plus a File System MCP server for internal Neo.ai.Agent local loops.

What is the Neo.mjs Agent OS?

The Neo.mjs Agent OS is the repository Brain: source code and services for Memory Core, Knowledge Base, Active Hybrid GraphRAG, DreamService, Golden Path synthesis, A2A coordination, and Neural Link tooling.

Frontmatter

id	15549
title	Measure zero-credential public-fleet first-paint budget
state	Closed
labels	documentationenhancementai
assignees	neo-gpt-emmy
createdAt	Jul 19, 2026, 1:12 AM
updatedAt	Jul 19, 2026, 6:30 AM
githubUrl	https://github.com/neomjs/neo/issues/15549
author	neo-gpt-emmy
commentsCount	0
parentIssue	15524
subIssues	[]
subIssuesCompleted	0
subIssuesTotal	0
contentTrust
projected
quarantined	0
signals	[]
blockedBy	[]
blocking	[]
closedAt	Jul 19, 2026, 6:30 AM

Measure zero-credential public-fleet first-paint budget

Closed Backlog/active-chunk-7 documentationenhancementai

neo-gpt-emmy commented on Jul 19, 2026, 1:12 AM

Context

This is an independently measurable child of #15524 under epic #15519 and the graduated authority in Discussion #15498 (OQ6/OQ7). The rate verdict is separately mergeable, while the source-mode transition and packaged witness remain sequenced behind #15543, #15544, and #15545.

This is not a micro-ticket: the result selects whether the flagship can honestly expose a zero-credential public-fleet mode or must keep public-fleet acquisition token-gated.

The Problem

The graduated plan requires a measured PASS/FAIL against GitHub's anonymous 60-request/hour REST budget before public-fleet wiring. The existing provider reader mixes REST, GraphQL, and trust-census operations. Without a live zero-token capability receipt, the product could expose a mode that cannot complete its own acquisition plan.

The Architectural Reality

ai/services/github-workflow/communityActivityShadowReader.mjs already owns GET-only acquisition and call/cost/gap receipts.
ai/services/github-workflow/probeCommunityActivityShadow.mjs binds its authenticated transport.
The tracked 30-day acquisition receipt records 1,714 provider request/cost units: 1,683 GraphQL and 31 REST.
A live token-cleared probe on 2026-07-18 observed anonymous REST limit 60, GraphQL limit 0 with HTTP 403, and HTTP 401 for the collaborators census. Public issue, issue-comment, and review-comment REST reads returned HTTP 200.
The correct placement is a sibling measurement under learn/agentos/measurements/; npm run --silent ai:structure-map -- --files --loc was run before authoring. No runtime owner or new .mjs file is introduced.

The Fix

Commit a versioned measurement that records the live zero-token capability receipt, the stable product verdict, and explicit reproduction triggers. The result must bind #15524 to:

bundled sample as the zero-call offline default;
token-present public fleet as an opt-in;
no second reader and no reduced REST subset relabeled as the full public fleet.

Contract Ledger

Authority / evidence	Consumer	Required effect
Live token-cleared provider receipt	Versioned measurement	Record capability facts, not a shared-machine remaining-counter policy
Existing reader request plan	Capability audit	Test the real plan; do not invent a second adapter
Discussion #15498 PASS/FAIL gate	#15524 implementation	Select the token-gated branch when anonymous capability is insufficient
Reproduction triggers	Future maintainer	Reopen only when provider capability or the canonical query plan changes

Decision Record Impact

Aligned with ADR 0037 / PR #15546. This measurement consumes that topology decision; it does not amend the topology.

Decision Record

No new ADR is needed. The ticket applies ADR 0037 and the graduated criterion from Discussion #15498.

Acceptance Criteria

A tracked measurement records the live zero-token REST, GraphQL, and collaborators results.
The verdict maps the observed capability failure to a token-gated public-fleet opt-in.
Volatile remaining counters are not promoted into product policy.
No production runtime is changed and no second provider reader is introduced.
Reproduction triggers are explicit.

Out of Scope

The bundled-sample to selected/live source-mode transition.
The packaged-demo witness.
Demo-host separation from #15545.
A new REST-only public preview.
Fleet Manager UI composition.

Avoided Traps

No fake zero-credential toggle whose canonical reader cannot complete.
No degraded REST subset presented as the full public fleet.
No throughput threshold inferred from a shared machine's current remaining counter.
No duplicated acquisition or fixture vocabulary.

Parent: #15524
Epic: #15519
Authority: Discussion #15498
Existing reader and receipt: #15149 / PR #15511
Topology ADR: PR #15546

Freshness Receipt

Immediately before creation at 2026-07-18T23:12:10.214Z, the latest 20 open issues, the latest 30 A2A messages across read states, and an exact/local resource sweep were checked. No equivalent ticket was found.

Handoff

The measurement is already implemented on codex/15524-fm-demo-authority; after this ticket is created, the commit and PR will be rebound to this child authority.

Origin Session ID: ad71d4c3-3e37-4a17-8df7-8415509def84