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	14573
title	Fleet cockpit PR and lane activity adapter
state	Closed
labels	enhancementaiarchitecture
assignees	neo-gpt
createdAt	3:24 AM
updatedAt	6:33 AM
githubUrl	https://github.com/neomjs/neo/issues/14573
author	neo-gpt
commentsCount	0
parentIssue	13015
subIssues	[]
subIssuesCompleted	0
subIssuesTotal	0
contentTrust
projected
quarantined	0
signals	[]
blockedBy	[]
blocking	[]
closedAt	6:33 AM

Fleet cockpit PR and lane activity adapter

Closed Backlog/active-chunk-2 enhancementaiarchitecture

neo-gpt commented on 3:24 AM

Context

Lane 2/C now has the cockpit DTO floor in PR #14571, and Ada explicitly split activity sources from fleet-wire in the #14561 decomposition sprint. #14572 covers A2A activity from Memory Core. This ticket covers the GitHub/PR/lane-state side of the cockpit activity feed.

The Problem

The Fleet cockpit needs to show what each agent is doing across PRs, issues, lane claims, and stalls. That data does not belong in FleetControlBridge and should not be guessed from UI text. Without a dedicated GitHub/lane adapter, Lane 1 can hardcode sample activity or couple to issue prose in a way that breaks as the Agent OS evolves.

The Architectural Reality

src/ai/fleet/fleetCockpitStatus.mjs defines the Body-side DTO and bounded event factory.
ai/services/graph/issueFocusSections.mjs already derives PR/participation/stall-oriented sections for agent workflow state.
ai/services/github-workflow/* owns GitHub issue/PR conversation and state reads.
apps/agentos/view/FleetSettingsPanel.mjs is only the current minimal consumer; #14560 owns the richer cockpit render.
Structure-map gate: ran npm run --silent ai:structure-map -- --files --loc; relevant owning folders are ai/services/graph, ai/services/github-workflow, src/ai/fleet, apps/agentos, and test/playwright/unit/ai/services/fleet.

The Fix

Add a small GitHub/lane activity adapter for the Fleet cockpit contract:

Read PR/issue/lane-state facts from the existing GitHub workflow / graph derivation surfaces.
Map them into bounded createFleetCockpitEvent() payloads with source labels and confidence.
Preserve agent identity, related ticket/PR ids, state class, and timestamps; avoid copying full public comments into the cockpit DTO.
Represent unavailable GitHub/graph reads as degraded/provisional capability state, not fake quietness.
Add focused unit coverage for mapping, redaction, missing-source behavior, and no sample data.

Contract Ledger Matrix

Target Surface	Source of Authority	Proposed Behavior	Fallback	Docs	Evidence
PR/lane cockpit events	`ai/services/github-workflow` + `issueFocusSections`	map current PR/issue/lane facts to source-labeled cockpit events	GitHub/graph unavailable -> degraded capability	issue body + source JSDoc	unit tests with PR/lane fixtures
Comment/body boundary	GitHub conversation state	keep ids, titles, states, timestamps; omit full comment bodies by default	redacted placeholder	source JSDoc	unit test proves body omission
Fleet DTO integration	`src/ai/fleet/fleetCockpitStatus.mjs`	emit bounded events that Lane 1 can render	adapter absent -> existing `not-wired` state	DTO JSDoc	unit test against DTO event factory

Decision Record impact

None. This is an adapter leaf inside the existing Fleet Manager / GitHub workflow / graph ownership split.

Acceptance Criteria

A GitHub/PR/lane activity adapter exists for Fleet cockpit consumption.
The adapter emits bounded, source-labeled cockpit events for PR, issue, lane-claim, and stall/progress facts where available.
Full comment bodies and secret-shaped fields are not exposed by default.
Missing GitHub/graph state is represented as degraded/provisional capability state, not fake empty activity.
Focused unit coverage proves mapping, bounds, redaction, and fail-closed behavior.

Out of Scope

A2A/Memory Core activity; tracked in #14572.
Runtime process truth; Ada owns the runtime-status wire method.
Lane 1 visual design/rendering in #14560.
Neural Link proof in #14563.

Avoided Traps

No fleet-wire method for GitHub activity.
No full comment-body stream into the cockpit DTO.
No UI/prose scraping as source of truth.
No sample activity in production paths.

#13015 · #14561 · #14562 · #14563 · #14571 · #14572 · #14560

Live latest-open sweep: checked latest 25 open issues immediately before creation at 2026-07-04T01:24Z after #14572 landed; no equivalent PR/lane activity adapter found. A2A in-flight sweep: checked recent mailbox claims; Ada requested this decomposition, no competing claim found. KB/local duplicate sweep: exact search returned no PR/lane activity adapter duplicate; Memory Core semantic search returned no existing activity-adapter ticket. Origin Session ID: 33403f62-0332-411a-bda3-0f4ab10cd1e6 Retrieval Hint: "Fleet cockpit PR lane activity adapter issueFocusSections GitHub workflow createFleetCockpitEvent"

tobiu referenced in commit 9edb342 - "feat(fleet): add PR/lane activity adapter (#14573) (#14630) on 6:33 AM

tobiu closed this issue on 6:33 AM