Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/50kg/zcode-deepseek-eye/setup-vision-helpernpx skills add 50kg/zcode-deepseek-eye --skill setup-vision-helpergit clone --depth 1 https://github.com/50kg/zcode-deepseek-eyeWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/50kg/zcode-deepseek-eye/setup-vision-helper)<a href="https://agentmods.dev/skills/50kg/zcode-deepseek-eye/setup-vision-helper"><img src="https://agentmods.dev/badge/skills/50kg/zcode-deepseek-eye/setup-vision-helper.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00106 | $0.01975 |
| Opus 5 | $0.00053 | $0.00988 |
| Sonnet 5 | $0.00021 | $0.00395 |
| Haiku 4.5 | $0.00011 | $0.00198 |
Grade A, and why
setup-vision-helper scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 6d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 111 lines — stays where its author put it; the contents beside it link to each section on GitHub.
配置视觉助手(借眼睛)
目标:让纯文本主模型(如 deepseek)在遇到图片时,自动通过一个视觉子 agent(vision-helper)"借眼睛";而当主模型本身支持视觉时,直接用主模型自己的眼睛,不调用子 agent。配置一次,之后切换主模型无需任何操作。
意图判断(先做)
本 skill 只会在两个明确表述下被触发,直接映射即可:
- 用户说"帮我配置视觉助手" → 走下方「执行步骤」配置流程。
- 用户说"删除视觉助手" / "卸载视觉助手" / "移除视觉助手" → 走下方「删除视觉助手」流程。
- 若触发了但用户表述与上述不符(例如只说了一半或加了奇怪内容)→ 先向用户确认意图,不要自行展开。
删除视觉助手
当用户意图是删除/卸载视觉助手时:
- 确认:先向用户确认"将删除 vision-helper 子 agent 和 AGENTS.md 中的路由规则,是否继续?"(删除不可逆)。
- 删除子 agent:删除
<主目录>/.zcode/agents/vision-helper.md(若存在)。 - 移除路由规则:从
<主目录>/.zcode/AGENTS.md中删除"图片处理(视觉助手)"小节(含vision-helper关键字的部分)。若删除后文件为空,可保留空文件或删掉文件,向用户说明。 - provider 配置默认不动:不要修改 config.json——补的
image模态(以及已有的apiFormat,如有)是该模型本身的正确配置(若它同时被当主模型用,删了反而会坏)。仅当用户明确要求回滚 config 时,才逐项询问要回滚哪一项。 - 报告:列出删除了什么、保留了什么(config 未动),提示重启 ZCode 后子 agent 才会从列表中消失。
重要原则(违反会出错)
- 不要用 modalities.input 过滤模型。ZCode 自定义模型的 modalities 字段经常滞后于真实能力(例如 qwen3-vl-flash 字段里只有 text,但实际支持图片)。候选清单列出所有已启用模型,让用户自己拍板哪个是"眼睛"。不要加提示、不要排除任何模型,只做清晰罗列。
- 不要向用户索要 API Key。key 应该由用户在 ZCode 设置里自行添加,本 skill 只复用已配置的 provider。
- 不要修改主模型选择,不要改动除"眼睛"以外的任何 provider/model 配置。
- 所有文件操作都用文件工具完成,并在结束时自检(文件存在、JSON 合法)。
执行步骤
1. 定位配置文件
- 解析用户主目录:
$HOME(macOS/Linux)或%USERPROFILE%(Windows),兜底os.homedir()。 - 主配置文件:
<主目录>/.zcode/v2/config.json。若不存在,搜索<主目录>/.zcode/下其他含provider键的 json(如config.json、cli/config.json),并告知用户实际使用的位置。 - 读取
provider对象。
2. 列出候选"眼睛"
- 收集所有
enabled(未显式禁用)的 provider 及其 models。 - 不加任何提示/注释,用编号列表清晰展示,格式:
[1] qwen3-vl-flash @ qwen3-vl-flash(deepseek的眼睛) [2] deepseek-v4-flash @ deepseek ... 请回复编号选择"眼睛"模型(或直接输入模型 id)。
3. 等待用户选择
- 用户回复编号 → 解析为对应模型;用户直接输入 id → 用之。
- 如果用户选择的是不在列表中的模型 id,允许,但提醒确认该模型已配置且可用。
4. 执行配置
a. 给眼睛模型补齐 image 模态(尽力而为,失败不中断):
- 在
provider.<该provider键>.models.<模型id>.modalities.input中,若不含"image",补上(["text","image"],保留原有其他值)。 - 若写入失败(如文件只读、被占用),警告用户可跳过——因为部分客户端不按该字段门控,补字段只是让 UI 与未来逻辑保持一致。
- 不要改动该模型的其他字段,不要改动其他任何模型。
b. 创建视觉子 agent(文件:<主目录>/.zcode/agents/vision-helper.md):
---
name: vision-helper
description: 看图并描述图片内容、回答与图片有关的问题。当主模型无法直接看到图像时使用。
model: custom:<眼睛模型所属 provider 的键>:<眼睛模型 id>
tools: [Read]
---
你是视觉助手。使用 Read 工具读取用户提供的图片文件路径,详细描述图片内容,
并回答主 agent 提出的关于这张图片的问题。最后用文字总结返回。
- model 字段必须用
custom:<provider键>:<模型id>格式(与"设置 → Subagents"界面生成的一致,provider 键即 config.json 中该 provider 的 key,例如custom:8b29396f-xxxx:qwen3-vl-flash)。不要用裸模型 id(会被当作默认 provider),也不要用手写斜杠格式。 - 若该文件已存在且
model与本次选择不同 → 先问用户"当前眼睛是 X,要换成 Y 吗?",确认后覆盖。 - 若已存在且相同 → 跳过,提示已配置。
- 提示用户:也可以直接在"设置 → Subagents"界面创建(Model 选眼睛模型),效果等同。
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 6d ago First seen · 111 lines · 106 tokens per session scan A 830fe1e089ec
setup-vision-helper is a skill published in the GitHub repository 50kg/zcode-deepseek-eye (41 stars, last pushed 25d ago), licensed MIT. It adds 106 tokens to every session and 1,975 once invoked, about $0.0005 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
agent-platform-rag-engine-management
Manage and query Agent Platform RAG Engine Corpora and retrieve grounded contexts using the Google GenAI SDK. Use when listing RAG corpora or files, inspecting a corpus, retrieving contexts, or generating content grounded in a RAG corpus. Do not use for standard database queries (use SQL/Spanner skills), Google…
agent-platform-model-registry
Agent Platform Model Registry Management. Use when you need to upload, list, describe, update, or delete machine learning models (and their versions) in the Agent Platform Model Registry. Don't use for model training, model deployment to endpoints, or managing non-Agent Platform models.
foundry-config-setup
Resolve missing setup caused by a hardcoded Foundry project endpoint or model in a sample. Use when a sample fails because it uses a placeholder/hardcoded projectendpoint (for example "https://your-project.services.ai.azure.com") or a hardcoded model instead of reading them from the environment.
google-cloud-solution-agentic-analytics-spark-knowledge-catalog
Discovers requirements and generates guidance to design and deploy a governed, secure agentic-analytics solution for data that's distributed across Google Cloud, other cloud providers, or on-premises. Data that's outside Google Cloud (such as data from Databricks, Snowflake, Salesforce, SAP, or Oracle systems) is…
training-check
Interactively monitor training metrics from the current Codex session, periodically checking WandB or fallback logs for NaN, divergence, plateaus, and broken runs.
nemo-automodel-launcher-config
Configure NeMo AutoModel job launches for interactive runs, Slurm clusters, and SkyPilot cloud execution.