fix(commandcode): share reasoning-facts table and fix GLM slug decoding for API-key preset

[youngchangjo]

Problem

The commandcode (API-key) registry entry was missing the official model-profile reasoning-facts table that the OAuth command-code entry carries. Two user-visible failures followed:

1. No reasoning picker + content.type 400 — the Codex catalog advertised supported_reasoning_levels: [] for every API-key model (deepseek-v4-flash/pro, GLM-5.x), so clients forced effort none and requests failed with:

   Provider error 400: messages.content.type is invalid, allowed values: ['text']
   

2. unsupported_model 400 — the router's known-ids decode source (knownModelIdsForProvider unions registry modelReasoningEfforts keys) missed the native slash ids, so Codex-facing slugs like commandcode/deepseek-deepseek-v4-pro were forwarded upstream verbatim (instead of deepseek/deepseek-v4-pro) and rejected intermittently (only when the live-models cache was cold):

   Provider error 400: Model "deepseek-deepseek-v4-pro" is not supported on this endpoint.
   

Fix

• src/providers/registry.ts: add modelReasoningEfforts: COMMAND_CODE_MODEL_REASONING_EFFORTS to the commandcode (API-key) entry — same shared table the OAuth command-code entry uses. This both restores the reasoning picker and feeds the router's known-ids decode source.
• src/providers/command-code-efforts.ts:
  • Fix GLM table keys to the exact upstream ids (zai-org/GLM-5.3, not zai-org/glm-5.3) so the exact-match known-ids decode works.
  • Make commandCodeReasoningEfforts / refreshCommandCodeReasoningEfforts lookups case-insensitive.
  • Add GLM-5, GLM-5.1, GLM-5.2-Fast (verified high/max from their official profile pages).

Verification

• New regression tests: API-key preset shares the OAuth reasoning-facts table; commandcode/deepseek-deepseek-v4-pro and commandcode/zai-org-GLM-5.3 decode back to native slash ids.
• bun test tests/slug-codec.test.ts tests/command-code-provider.test.ts tests/routing-compatibility.test.ts tests/provider-registry-parity.test.ts — all pass.
• Live verification against the real Command Code provider API after applying the same change locally:
  • Catalog now advertises deepseek-* [high, max, ultra], zai-org-GLM-5.3 [low, high, max, ultra].
  • commandcode/deepseek-deepseek-v4-pro and commandcode/zai-org-GLM-5.3 requests (incl. reasoning_effort) return 200 with reasoning tokens.

Review readiness checklist

This PR stays in draft until every box below is ticked. Tick all four boxes once the requirements are met:

• All CI tests are green on my local testing.

• I pushed my PR to the latest dev commit.

• I resolved all correct Codex and CodeRabbit findings.

• My PR is ready for review.

Summary by CodeRabbit

• New Features

  • Added support for GLM-5, GLM-5.1, GLM-5.2-Fast, and GLM-5.3 models.
  • Added official reasoning-effort support for Command Code models.
  • Improved support for namespaced models, including IDs containing slashes.

• Bug Fixes

  • Model IDs are now matched case-insensitively for more reliable lookups.
  • Corrected canonical model ID casing and alias decoding.

[github-actions]

✅ Deterministic PR hygiene checks passed.

[coderabbitai]

No actionable comments were generated in the recent review. 🎉

ℹ️ Recent review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: 62293da3-986f-43d6-9e53-6fa9816ad92c

📥 Commits

Reviewing files that changed from the base of the PR and between 2a1604e and 011b11d.

📒 Files selected for processing (4)

• src/providers/command-code-efforts.ts
• src/providers/registry.ts
• tests/command-code-provider.test.ts
• tests/slug-codec.test.ts

Included review availability: Your plan includes up to 10 reviews per rolling hour; 9 remain after this review.

---

📝 Walkthrough

Walkthrough

The change aligns Command Code GLM model identifiers with upstream casing, adds new GLM profiles, makes profile lookups case-insensitive, shares reasoning metadata with the API-key provider, and adds registry and alias regression coverage.

Changes

Command Code model support

Layer / File(s)	Summary
Model profiles and case-insensitive lookup src/providers/command-code-efforts.ts	Adds GLM-5, GLM-5.1, and GLM-5.2-Fast profiles. Updates GLM identifiers to upstream casing. Reasoning-effort and profile lookups now match configured model IDs case-insensitively.
Registry wiring and regression coverage src/providers/registry.ts, tests/command-code-provider.test.ts, tests/slug-codec.test.ts	The API-key provider now uses shared Command Code reasoning metadata. Tests cover canonical model IDs, registry parity, native slash-containing IDs, and routed alias decoding.

Estimated code review effort: 3 (Moderate) | ~20 minutes

Merge Risk: ⚪ Minimal · up to 011b1

The change restores reasoning options and correct model-id decoding for the API-key provider, with regression coverage and passing verification; no actionable merge-blocking risk remains beyond normal checks and review.

Possibly related PRs

• lidge-jun/opencodex#923: Introduces related Command Code provider registry and model-ID handling.
• lidge-jun/opencodex#1066: Extends related Command Code reasoning metadata and registry integration.
• lidge-jun/opencodex#1197: Updates related GLM and routed-model reasoning metadata.

Suggested reviewers: ingwannu, wibias

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Linked Issues check	✅ Passed	The changes satisfy the provider routing, reasoning-effort mapping, model synchronization, and regression coverage objectives in [#39].
Out of Scope Changes check	✅ Passed	All changed files support the stated Command Code reasoning-table, model-casing, slug-decoding, and regression-test objectives.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Description Check	✅ Passed	Check skipped - CodeRabbit’s high-level summary is enabled.
Title check	✅ Passed	The title clearly identifies the main changes: shared reasoning data and corrected GLM slug decoding for the Command Code API-key preset.

✨ Finishing Touches 💡 1

🛠️ Fix failing CI checks 💡

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[github-actions]

⏳ DRAFT

• review readiness checklist open (0/4 boxes ticked).

What to do

• Tick all four boxes in the PR description once you're done (currently 0/4).

Review readiness checklist

• ⬜ All CI tests are green on my local testing.
• ⬜ I pushed my PR to the latest dev commit.
• ⬜ I resolved all correct Codex and CodeRabbit findings.
• ⬜ My PR is ready for review.

0/4 boxes ticked.

This PR stays in draft until every box above is ticked.

[Wibias]

@youngchangjo please retarget this pr on the dev branch.

[lidge-jun]

Retargeted base from main to dev per branch policy: main only moves by maintainer promotion, and this head (011b11d) is not part of a release train. The 4-file commandcode reasoning/slug fix still applies on dev (modelReasoningEfforts absent, GLM keys still glm-5.3). Needs a rebase before review.

[lidge-jun]

Landed via #2015 — your commits merged onto current dev with the one slug-codec test conflict resolved additively (both suites kept; 89/0 + tsc clean). Thanks for the fix: the commandcode API-key preset now shares the reasoning-facts table and GLM slugs decode with canonical casing. Closing this original since the work is on dev.