Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →lingtai-ai/preset-skill-deepseek
Official-source-led manual for the TUI `deepseek` template.
Passed 19 of 22 test cases with the skill, and 5 of 22 without it.
and 21 more in the full report
14 fixed5 passed either way3 still failing0 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 14 test cases the skill fixed. Same model, same prompt, with and without the skill.
Where in our TUI Go repository is the built-in deepseekPreset struct defined, and what is its default model ID? We are tempted to look in configs/presets.json or internal/models/deepseek.go assuming standard model names like deepseek-chat.
| Checked | Without skill | With skill |
|---|---|---|
| identifies tui/internal/preset/preset.go as the definition location | Missed | Passed |
| specifies deepseek-v4-pro as the default model | Missed | Passed |
The instructions your model reads when the skill loads.
Use this child for the named built-in deepseek preset. deepseekPreset in tui/internal/preset/preset.go:1321 uses provider deepseek, model deepseek-v4-pro, https://api.deepseek.com, DEEPSEEK_API_KEY, OpenAI compatibility, web_search, and skills. It has no built-in vision capability. The editor also offers DeepSeek API, OpenCode Go, and Custom base_url rows.
Read the official DeepSeek models and pricing and models API. For the native route, authenticated GET https://api.deepseek.com/models is the served-list check. OpenCode Go is a separate gateway: query its authenticated GET https://opencode.ai/zen/go/v1/models and keep only DeepSeek IDs served there; do not use the native list as proof for Go. Preserve the route-specific credential and model spelling.
The picker currently carries deepseek-v4-pro and deepseek-v4-flash, both modelHasVision false. The experimental deepseek-v4-flash-vision-exp entry is not part of the stock picker or vision contract. OpenCode Go is scoped to DeepSeek IDs for this provider; other Go models belong in Custom. The native API row and Go row select DEEPSEEK_API_KEY and OPENCODE_GO_API_KEY respectively; an edited native built-in receives a numbered DEEPSEEK_1_API_KEY-style slot.
Start at deepseekPreset in tui/internal/preset/preset.go. Revise providerModels["deepseek"] and its modelHasVision entries in tui/internal/tui/preset_editor.go for picker or vision changes. Revise ProviderRegionURLs and ProviderDefaultEnv for base_url/env behavior; the constructor remains text-only unless a reviewed native vision route is wired. Follow tui/internal/tui/SKILL.md and tui/CONTRACT.md.
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, use dry-run/check before apply, and
test cases passed, vs 5 / 22 without
Published skills whose descriptions closely match this one, with their headline benchmark lift.
More skills from the same source and nearby in the registry, with their headline benchmark lift.