基于 SOC 职业分类
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/Lingtai-AI/lingtai --skill preset-skill-zhipu命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
正在显示 SKILL.md
| name | preset-skill-zhipu |
| description | Use when revising the built-in zhipu TUI preset. |
| version | 3.0.0 |
| last_changed_at | 2026-09-07T00:00:00Z |
| related_files | ["tui/internal/preset/preset.go","tui/internal/tui/preset_editor.go","tui/internal/tui/SKILL.md","tui/CONTRACT.md","tui/internal/preset/revision.go","tui/internal/headless/preset_revision.go"] |
| maintenance | If you find stale or incorrect information here, use the lingtai-issue-report skill to assemble evidence and obtain per-issue human consent before filing an issue. Never include secrets, credentials, tokens, or private paths. |
Use this child for the named built-in zhipu preset. zhipuPreset in tui/internal/preset/preset.go:1275 ships GLM-5.2, ZHIPU_API_KEY, OpenAI compatibility, CN/INTL coding endpoints, web_search, and skills; it does not ship a vision capability. ProviderRegionURLs["zhipu"] also has an OpenCode Go row whose model spelling is lowercase.
Read the official Zhipu model overview and Z.AI model docs for the applicable region. For the native CN/INTL route, use that provider catalog and its documented API model list/endpoint; for OpenCode Go, separately call the gateway's authenticated GET https://opencode.ai/zen/go/v1/models and preserve the lowercase IDs it serves. Do not treat GLM-4.6V or the optional Z.AI Vision MCP as the text preset's native vision. Preserve the exact case required by the selected route. The optional official server uses package @z_ai/mcp-server and backing model GLM-4.6V; it reads Z_AI_API_KEY and Z_AI_MODE, while this preset reads ZHIPU_API_KEY. Vision Understanding consumes a separate 5-hour prompt pool; never install, register, or invoke that MCP automatically.
Start at zhipuPreset in tui/internal/preset/preset.go. Revise providerModels["zhipu"] and all matching modelHasVision entries in tui/internal/tui/preset_editor.go when the picker or text/vision fact changes. For endpoint or credential changes, revise ProviderRegionURLs and ProviderDefaultEnv in preset.go. Recheck the constructor's intentionally absent vision key and the model-row/base_url behavior; follow tui/internal/tui/SKILL.md and tui/CONTRACT.md.
The picker retains uppercase GLM-5.2/GLM-5.1 and lowercase glm-5.2/glm-5.1 in one mixed list. This is an evidence-backed no-op: the reviewed sources do not pin a native Coding Plan GLM-5.3 ID or casing strongly enough to split the native and Go routes safely, so no native GLM-5.3 entry is invented. The constructor's absent vision key is intentional: the shipped GLM coding models are text-only.
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 apply only to a new explicit output directory. revision.go validates hashes, route bindings, and evidence and preserves unowned JSON bytes. This amendment records the native-picker no-op and preserves the current constructor, Go behavior, endpoint, credential, and capability values.
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.
For save, endpoint/capability, availability, activation/refresh, or
troubleshooting, use the five shared operation children under
reference/operations/; this child owns Zhipu route, model, and vision facts.