How this fits together

The model-list CLI reads models and discovery outcomes published by the selected Gateway or local catalog owner. It presents these facts as a table, JSON, or plain model keys.

flowchart TD
 A[User lists models] --> B[Selected Gateway or local catalog]
 B --> C[Published models and discovery outcomes]
 C --> D[Present discovery diagnostics]
 D --> E[Stderr with recovery guidance]
 C --> F[Render requested output]
 F --> G[Table, JSON, or model keys]

Technical review

Best possible solution:

Present the catalog owner's existing failure facts at the CLI boundary and investigate the remaining xAI entitlement and routing claims separately.

Do we have a high-confidence way to reproduce the issue?

Yes: an auth-rejected provider outcome without refreshFailed deterministically receives no provider diagnostic on current main. Source and existing coverage establish the trigger; this review did not execute code.

Is this the best way to solve the issue?

Yes for the diagnostic gap: displaying already-published facts is a narrow repair that preserves discovery ownership and avoids speculative authentication or routing changes.

AGENTS.md: found and applied where relevant.

Codex review notes: model internal, reasoning medium; reviewed against 5e8fbaf8b422.

Provenance checked

• src/commands/models/list.list-command.ts: published-catalog consumption keeps the original intent (refactor(models): read published catalog inventory through the Gateway #142063: Read the selected Gateway's published inventory without competing local discovery or secret resolution; explicit refresh owns acquisition.)
• src/commands/models/list.list-command.ts, docs/cli/models.md, and prepared-catalog coverage: rejection classification keeps the original intent (fix(models): separate rejected auth from catalog refresh failures #158207: Treat rejected authentication as provider state rather than a transient refresh failure, and consume the publisher's refreshFailed fact.)
• src/commands/models/list.table.ts: JSON output keeps the original intent (4ee41cc: Separate JSON payload output from logging.)

Testing

Proof path: shipped entry point. Added test files: 1.

Security

None.

Evidence

What I checked:

• Introduced presentation repair: The pinned introduced diff adds diagnostics for existing provider outcomes and preserves the publisher-owned refreshFailed flag, model filtering, discovery, and routing. (src/commands/models/list.list-command.ts:138, 1d4a7a9b41e2)
• Current-main necessity: Fetched main still prints only the generic refreshFailed warning and model rows; it does not present providerOutcomes. The touched production files, documentation, and regression suite have no changes between the merge base and fetched main, so this diagnostic repair is not already implemented. (src/commands/models/list.list-command.ts:133, 5e8fbaf8b422)
• Public outcome contract: The existing Gateway schema already defines provider, optional profileId, and ready/auth-rejected/unavailable status. The CLI JSON projection selects only these public fields and does not expose raw provider error bodies. (packages/gateway-protocol/src/schema/model-catalog.ts:125, 1d4a7a9b41e2)
• Intentional rejection classification: Merged fix(models): separate rejected auth from catalog refresh failures #158207 intentionally separated rejected authentication from refresh failure. Its source patch removes duplicate CLI classification; this PR retains that decision while displaying a separate diagnostic. (src/commands/models/list.list-command.ts:133, 3685f7a7fb22)
• Catalog ownership rationale: Merged refactor(models): read published catalog inventory through the Gateway #142063 made published inventory authoritative and reserved acquisition for explicit refresh. Its full PR body was read through the GitHub REST API; this patch preserves that boundary. (src/commands/models/list.list-command.ts:80, 89fcceb655c1)
• After-fix shipped CLI proof: The supplied full PR body, captured as sourceRevision bd1d55332da6815c11b0456afb6f33b4945dad4d06d7723a3e42237db134376b, reports five built-CLI scenarios at the reviewed head on macOS arm64 against an isolated WebSocket catalog fixture. Human, JSON, and plain empty catalogs emitted authentication diagnostics with exit 0; retained JSON/plain catalogs preserved catalog-provider/Reader and machine-readable stdout. This exercises CLI presentation rather than live xAI entitlement. (1d4a7a9b41e2)

Likely related people:

• Peter Steinberger: Raw commit 3685f7a adds src/commands/models/list.list-command.ts:133 relative to its recorded parents. This identifies author metadata, not feature responsibility or a PR merger. (role: source-line author; confidence: high; commits: 3685f7a7fb22; files: src/commands/models/list.list-command.ts)
• unknown: The claimed source-line change could not be verified from bounded local history. (role: source history unknown; confidence: low)

PR surface

Source +25, Tests +79, Docs +9. Total +113 across 4 files.

Root-cause cluster

Relationship: fixed_by_candidate
Canonical: #166875
Summary: This PR is a candidate repair for the linked issue's diagnostic omission, not its broader entitlement and routing claims.

Members:

• canonical: xAI OAuth (SuperGrok/Premium+): recommended path silently yields no usable models — catalog discovery rejected by cli-chat-proxy; acpx grok-build harness works #166875 - The linked report establishes the silent catalog-rejection problem and tracks the remaining provider claims.

Proposal only: this assessment does not dispatch repair, suppress jobs, mutate sibling items, close, or merge anything.

Labels

Label changes:

• add rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🐚 platinum hermit and patch quality is 🐚 platinum hermit.
• remove rating: 🐚 platinum hermit: Current PR rating is rating: 🦐 gold shrimp, so this older rating label is no longer current.

Label justifications:

• P2: This is a bounded CLI diagnostic repair that does not change provider entitlement, authentication, or routing.
• rating: 🦐 gold shrimp: Overall readiness is 🦐 gold shrimp; proof is 🐚 platinum hermit and patch quality is 🐚 platinum hermit.
• status: 👀 ready for maintainer look: ClawSweeper has no concrete contributor-facing blocker left for this PR.
• proof: sufficient: Contributor real behavior proof is sufficient.

Rating scale

6/6 🦀 challenger crab · 5/6 🦞 diamond lobster · 4/6 🐚 platinum hermit · 3/6 🦐 gold shrimp · 2/6 🦪 silver shellfish · 1/6 🧂 unranked krab. Overall follows the weaker of proof and patch quality; ✨ marks media proof (a screenshot, video, or linked artifact) that directly shows the changed behavior.

Workflow

ClawSweeper edits this one comment on every review. Comment @clawsweeper re-review for a fresh review only; repair and merge need explicit maintainer commands such as @clawsweeper autofix or @clawsweeper automerge.

History