Install any skill in seconds. Free to start, no credit card required.
Get Started Free →Official-source-led manual for the TUI `openrouter` template.
.claude/skills/lingtai-ai-preset-skill-openrouter/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | -22% | 0% |
| case-02 | ✗→✓ | ▲ Improved | -41% | 0% |
| case-03 | ✗→✓ | ▲ Improved | -31% | 0% |
| case-04 | ✗→✓ | ▲ Improved | -83% | 0% |
| case-06 | ✗→✓ | ▲ Improved | -68% | 0% |
openrouteropenrouterPreset() in tui/internal/preset/preset.go ships gateway model z-ai/glm-5.1, provider openrouter, a provider-resolved base URL, and OPENROUTER_API_KEY. The exact shipped slug is text-in/text-out, and the manifest has no vision capability.
The stock template is text-only and does not invoke the vision tool. If an operator creates a user-owned saved preset that explicitly adds capabilities.vision, actual image support still depends on the selected downstream model; only then may the vision tool attempt the saved preset's endpoint, model, and credential over the compatible route. Gateway-wide multimodality alone is not proof that a particular downstream model accepts images.
Read the official OpenRouter model guide and image-understanding guide on demand. No OpenRouter plan-level vision MCP is evidenced. If the real image request fails, the sanitized vision tool result reports the failure type and points to vision(action="manual") for explicit alternatives. Do not silently change model/provider or auto-load/invoke an MCP. Verify provider, model, endpoint resolution, and capability flags in TUI source.
For base URL/API-compat/model/capability declaration shape versus credentials, see reference/operations/endpoint-capabilities/SKILL.md.
Other measured skills in the registry, with their headline benchmark lift.