#20136: Telegram: fix cross-provider model checkmark in /models keyboard

## Summary - Problem: In Telegram model keyboard (`/models`), active checkmark used only model id and ignored provider. - Why it matters: Same model id under different providers (for example `openai-codex/gpt-5.3-codex` and `github-copilot/gpt-5.3-codex`) showed a false active `✓`. - What changed: `buildModelsKeyboard` now parses `currentModel` as `provider/model` and marks active only when both provider and model match; added regression test for cross-provider duplicate model ids. - What did NOT change (scope boundary): Model selection flow, callback handling, and session override persistence logic were not changed. ## Change Type (select all) - [x] Bug fix - [ ] Feature - [ ] Refactor - [ ] Docs - [ ] Security hardening - [ ] Chore/infra ## Scope (select all touched areas) - [ ] Gateway / orchestration - [ ] Skills / tool execution - [ ] Auth / tokens - [ ] Memory / storage - [x] Integrations - [ ] API / contracts - [x] UI / DX - [ ] CI/CD / infra ## User-visible / Behavior Changes - In Telegram model picker, `✓` is now shown only for the exact active `provider/model` pair. - No behavior change in actual model switching command execution. ## Security Impact (required) - New permissions/capabilities? `No` - Secrets/tokens handling changed? `No` - New/changed network calls? `No` - Command/tool execution surface changed? `No` - Data access scope changed? `No` - If any `Yes`, explain risk + mitigation: `N/A` ## Repro + Verification ### Environment - OS: macOS - Runtime/container: Node `v22.19.0` + pnpm `10.23.0` (no container) - Model/provider: `openai-codex/gpt-5.3-codex`, `github-copilot/gpt-5.3-codex` - Integration/channel (if any): Telegram - Relevant config (redacted): allowlist includes both providers with overlapping model id ### Steps 1. Set current session model to `openai-codex/gpt-5.3-codex`. 2. Open Telegram `/models github-copilot`. 3. Observe model row `gpt-5.3-codex`. ### Expected - No `✓` on `github-copilot/gpt-5.3-codex` when active model is `openai-codex/gpt-5.3-codex`. ### Actual - Before fix: `✓` was shown incorrectly. - After fix: `✓` is shown only on exact provider/model match. ## Evidence - [ ] Failing test/log before + passing after - [x] Trace/log snippets - [ ] Screenshot/recording - [ ] Perf numbers (if relevant) `pnpm exec vitest run src/telegram/model-buttons.test.ts` Result: `✓ src/telegram/model-buttons.test.ts (29 tests)` ## Human Verification (required) - Verified scenarios: Added and ran regression test for duplicate model id across different providers; confirmed no checkmark on wrong provider page. - Edge cases checked: Existing test for same-provider active checkmark still passes; pagination/callback-size tests still pass. - What you did **not** verify: Manual live Telegram interaction against a running bot instance. ## Compatibility / Migration - Backward compatible? `Yes` - Config/env changes? `No` - Migration needed? `No` - If yes, exact upgrade steps: `N/A` ## Failure Recovery (if this breaks) - How to disable/revert this change quickly: Revert this commit. - Files/config to restore: `src/telegram/model-buttons.ts`, `src/telegram/model-buttons.test.ts` - Known bad symptoms reviewers should watch for: Missing `✓` for active model when provider aliases differ unexpectedly. ## Risks and Mitigations - Risk: If provider alias forms differ (`z.ai` vs `zai`) in `currentModel` vs keyboard provider id, `✓` may be absent. - Mitigation: Selection/switch logic is unaffected (still uses full `/model provider/model` path); this change affects only visual active marker. <!-- greptile_comment --> <h3>Greptile Summary</h3> Fixes a bug in the Telegram `/models` keyboard where the active model checkmark (`✓`) was shown based solely on model ID, ignoring the provider. This caused false-positive checkmarks when different providers offered models with the same ID (e.g., `openai-codex/gpt-5.3-codex` vs `github-copilot/gpt-5.3-codex`). - `buildModelsKeyboard` now parses `currentModel` into both provider and model ID, and requires both to match before showing the `✓` marker - Backward-compatible: when `currentModel` has no provider prefix, the function falls back to model-ID-only matching - Added a regression test for the cross-provider duplicate model ID scenario <h3>Confidence Score: 5/5</h3> - This PR is safe to merge — it's a focused, well-tested bug fix with no behavioral side effects outside the visual checkmark logic. - The change is small, targeted, and correct. It fixes a clear bug (cross-provider false-positive checkmark) with proper backward compatibility for provider-less model strings. The parsing logic correctly handles edge cases (nested model paths, empty/whitespace strings). Both callers in bot-handlers.ts and commands-models.ts pass the same provider/currentModel shape. A regression test is included. - No files require special attention. _{Last reviewed commit: 0e10d63} <!-- greptile_other_comments_section --> _{(3/5) Reply to the agent's comments like "Can you suggest a fix for this @greptileai?" or ask follow-up questions!} <!-- /greptile_comment -->