lingtai-ai minimax preset revision
Use this child for the named built-in minimax preset. Its current constructor is minimaxPreset in tui/internal/preset/preset.go:1248: it ships MiniMax-M2.7, MINIMAX_API_KEY, the CN Anthropic-compatible default, and text/tool capabilities. The regional rows are ProviderRegionURLs["minimax"]: CN, INTL, and OpenCode Go.
Template-specific settings
Read the official MiniMax Models catalog and its API overview. For served IDs, call the selected official Anthropic route's authenticated GET /anthropic/v1/models; use the catalog/API docs for modality and plan facts. OpenCode Go is a gateway, so separately query its authenticated GET https://opencode.ai/zen/go/v1/models and do not copy a MiniMax catalog into the Go row. Preserve exact case and distinguish the Go credential from the regional slots. Check the official Token Plan MCP guide only for its separate manual image tool; it does not make the TUI vision capability change. The official MiniMax-Coding-Plan-MCP repository ships the minimax-coding-plan-mcp package and understand_image tool for a Token Plan seat or purchased Credits; it is manual-only and not auto-wired here.
TUI surfaces to revise
Start at minimaxPreset in tui/internal/preset/preset.go. For a model change, revise providerModels["minimax"] and the matching modelHasVision entries in tui/internal/tui/preset_editor.go; for endpoint or credential changes, revise ProviderRegionURLs and ProviderDefaultEnv in preset.go. Confirm the fixed capability rendering only if the constructor's vision wiring changes. Follow the curation and retired-model rules in tui/internal/tui/SKILL.md and tui/CONTRACT.md.
The native CN picker is MiniMax-M2.7, MiniMax-M2.7-highspeed, MiniMax-M2.5, and MiniMax-M2.5-highspeed, all explicitly text-only for the reviewed Anthropic-compatible route. INTL remains free text because native parity is not proven. OpenCode Go keeps its exact pre-PR picker — MiniMax-M3, MiniMax-M2.7, and MiniMax-M2.7-highspeed — and its prior behavior; native modality metadata is not copied into that gateway route. Retired IDs remain valid only when already pinned in a saved preset.
Reviewed deterministic revision
Prepare an evidence-bound manifest and explicit input, then run lingtai-tui presets revise --manifest PATH --input PATH --mode dry-run|check|apply [--output-dir PATH]. Review the JSON plan and diagnostics, run dry-run/check before apply, and apply only to a new explicit output directory. The adapter reads only those paths; revision.go validates hashes, route bindings, and evidence and preserves unowned JSON bytes. This amendment revises native CN curation while preserving the OpenCode Go route and all endpoint/credential semantics.
Maintenance: If the relevant TUI preset/page is revised, check whether this sub-skill also needs revision and, if so, include it in the same PR.
Operations
For save, endpoint/capability, availability, activation/refresh, or
troubleshooting, use the five shared operation children under
reference/operations/; this child owns MiniMax route, model, and vision
facts.