feat: add cerebras/zai-glm-4.7 parameters

[guillaumegay13]

Adds parameter coverage for cerebras/zai-glm-4.7 (Cerebras-served GLM-4.7). Addresses #85.

New provider cerebras (api_key), plus a lobe-icons logo mapping.

Parameters (10)

max_completion_tokens, temperature, top_p, frequency_penalty, presence_penalty, seed, stop, reasoning_effort, clear_thinking, response_format.type.

Verified against the live Cerebras API

Every parameter was smoke-tested against POST /v1/chat/completions with model: zai-glm-4.7. Each returned 200; an unsupported param like top_k returns 400, so acceptance is real.

• reasoning_effort values are none / low / medium / high — confirmed by the API's own validation error: "Input should be 'none', 'low', 'medium' or 'high'" (default is rejected, unlike Groq's build of the same knob).
• clear_thinking (boolean, default true) is GLM-4.7-specific — controls whether prior-turn thinking is kept in context.

Validated locally: npm run validate, npm test, npm run guard:params, npm run typecheck, npm run build all pass.

[vercel]

@guillaumegay13 is attempting to deploy a commit to the Manifest Team on Vercel.

A member of the Team first needs to authorize it.