PR Review Summary

Status: Drop+Supersede

🪜 Strategic-Fit Decision

Per §9 Strategic-Fit Step-Back:

• Decision: Drop+Supersede

• Rationale: The source ticket's latent defect does not exist in Neo's runtime object model. FleetManager.mjs exports the singleton instance returned by Neo.setupClass(FleetManager); the producer and consumers all operate on that same instance. Moving the declarations to constructor statics separates them from the live writes instead of aligning them.

• Disposition: ticket-premise-dead

• Source-coordinate falsifiers: ai/services/fleet/FleetManager.mjs ends with export default Neo.setupClass(FleetManager); src/Neo.mjs:993-997 instantiates and exports singleton classes; ai/services/fleet/devFleetServer.mjs:144-186 writes the imported singleton; ai/services/fleet/FleetControlBridge.mjs:202-204 returns that same singleton. At exact head, Neo.typeOf(FleetManager) === 'NeoInstance'; the three properties are absent from the export and present only on FleetManager.constructor. After the production-shaped assignment, the export owns the value while the constructor remains null.

• Salvage map: Preserve the exact instance-vs-constructor runtime census as the closure evidence. Discard the three constructor-static moves and both new tests; they encode the inverted identity model.

• Successor landing pad: None — close #16786 as premise-dead. There is no remaining defect to route into a new ticket.

• Successor map citation: #16786 plus this review's source-coordinate and salvage map; no code successor is warranted.

Peer-Review Opening: Phoebe — the instinct to make the declaration match the write was good. The singleton boundary reverses the ticket premise, though, and the two red tests are correctly exposing that mismatch.

🧭 Patch-Blind Premise Snapshot

• Inputs Read Before Patch: #16786; the two-file changed-file list; current dev FleetManager; Neo.setupClass / Neo.create; the live devFleetServer writer; the FleetControlBridge consumer; three Memory Core prior-art searches.
• Expected Solution Shape: A declaration/write mismatch must be repaired on the exact runtime object both sides use. For a Neo singleton, a discriminating test must first identify whether the module export is the constructor or the instantiated singleton.
• Patch Verdict: Contradicts the expected shape. Base already declares all three fields on the exported singleton instance. The patch moves them to a constructor object neither the writer nor readers use.
• Premise Coherence: Conflicts with verify-before-assert: the ticket and patch infer constructor-static authority without executing Neo.setupClass's singleton return boundary. Flat-peer and two-hemisphere values are otherwise N/A to this narrow service-internal change.

🕸️ Context & Graph Linking

• Target Epic / Issue ID: Resolves #16786
• Related Graph Nodes: #15271 · #15272 · #15273 · FleetManager · class-as-singleton injection
• Origin Session ID: 878f05af-2c4e-4da2-a5c2-9e4af666fcb8

🔬 Depth Floor

Challenge: Execute the exact module export rather than reading the class declaration in isolation. Exact-head results:

• Neo.typeOf(FleetManager) is NeoInstance;
• Object.hasOwn(FleetManager, key) is false while Object.hasOwn(FleetManager.constructor, key) is true;
• FleetManager.wakeStateOptions = value creates the property on the exported instance and leaves the constructor static null;
• Neo.create(FleetManager) rejects because its argument is already a Neo instance;
• Neo.create(FleetManager.constructor) creates a different instance that does not see the live singleton's injected value.

The base control shows the inverse and correct topology: all three declarations are own properties of the exported singleton, and the production-shaped assignment overwrites that same property.

Rhetorical-Drift Audit:

• PR description: the claimed “CLASS-static write” is actually a write to the exported singleton instance.
• Anchor & Echo summaries: the new JSDoc repeats the same false constructor-static model.
• [RETROSPECTIVE] tag: N/A — absent.
• Linked anchors: the sanctioned seam tickets do not establish that the module export is a constructor.

Findings: Required drift is structural, not editorial: the prose describes an object identity the runtime falsifies.

🧠 Graph Ingestion Notes

• [KB_GAP]: For a class with singleton:true, Neo.setupClass returns and registers the instantiated singleton. A JavaScript static field remains on its constructor and is not the declaration site for writes to the exported instance.
• [TOOLING_GAP]: None. Hosted unit CI and the exact-head focused run both caught the premise inversion.
• [RETROSPECTIVE]: Before class-vs-instance injection hygiene work, execute one identity census: Neo.typeOf(export), own keys on export, own keys on export.constructor, then the production-shaped write.

N/A Audits — 📑 🪜 📡 🔗

N/A across listed dimensions: no public contract ledger, externally observable evidence ceiling, OpenAPI surface, skill, or cross-substrate convention is changed.

🎯 Close-Target Audit

• Close-targets identified: #16786
• #16786 confirmed not epic-labeled.

Findings: The mechanics pass, but closing through this implementation would falsely record a non-defect as fixed; close the ticket as premise-dead instead.

🧪 Test-Evidence & Location Audit

• Execution evidence: exact-head required CI is red at 7c3cf69bccfc63fac8c1a07e3bbc9c0fe483b331; the two newly added FleetManager.spec tests fail.
• Reviewer falsifier: npm run test-unit -- ai/services/fleet/FleetManager.spec in an exact-head snapshot reproduced 2 failures / 18 passes. A direct runtime census then distinguished exported instance from constructor and compared the base control.
• Test location: the file location is correct; the asserted object identity is not.

Findings: Required evidence fails in the exact semantic surface this PR introduces. The failure is causal, not ambient CI noise.

📋 Required Actions

To proceed with closure:

• Close this PR without merge and close #16786 as premise-dead; no replacement implementation or new follow-up ticket is warranted.

📊 Evaluation Metrics

• [ARCH_ALIGNMENT]: 10 - Constructor statics are placed on the wrong runtime object for this singleton seam.
• [CONTENT_COMPLETENESS]: 20 - The diff, tests, and prose are internally consistent, but all encode the inverted singleton identity.
• [EXECUTION_QUALITY]: 10 - Both new exact-head tests fail and the direct runtime probe falsifies the claimed write path.
• [PRODUCTIVITY]: 5 - Iterating this implementation would spend review cycles on a non-defect.
• [IMPACT]: 5 - No production failure is fixed; merge would mainly institutionalize false JSDoc.
• [COMPLEXITY]: 15 - The diff is small, but its apparent simplicity hides the class/instance boundary.
• [EFFORT_PROFILE]: Maintenance - closure and evidence folding, not another code cycle.

The clean outcome is to preserve the falsifier and remove the lane, not to repair the red tests around a dead premise.

[review-budget-managed]

• outcome: terminal-drop-supersede
• ordinary-limit: 2
• activation-issue: 15257
• activation-pr: 15307
• activated-at: 2026-07-16T20:54:31Z