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 skills add LINoone516/lino-agent-skills --skill lino-local-visiongit clone --depth 1 https://github.com/LINoone516/lino-agent-skillsWrote 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/linoone516/lino-agent-skills/lino-local-vision)<a href="https://agentmods.dev/skills/linoone516/lino-agent-skills/lino-local-vision"><img src="https://agentmods.dev/badge/skills/linoone516/lino-agent-skills/lino-local-vision/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/linoone516/lino-agent-skills/lino-local-vision"><img src="https://agentmods.dev/badge/skills/linoone516/lino-agent-skills/lino-local-vision.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00057 | $0.01243 |
| Opus 5 | $0.00028 | $0.00622 |
| Sonnet 5 | $0.00011 | $0.00249 |
| Haiku 4.5 | $0.00006 | $0.00124 |
Grade A, and why
lino-local-vision 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 11d 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 — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Local Vision(LM Studio 本地视觉模型)
调用本机 LM Studio 的本地多模态模型分析图片,纯本地、不走网络、不需要 API key、无费用。默认模型 qwen3-vl-8b-instruct(官方指令版,无思维链、直接输出,识别更准);备选 qwen3.5-9b-uncensored-hauhaucs-aggressive(第三方社区模型,带思维链需压 reasoning,可在 LM Studio 的 Model 搜索里找到,或从 ModelScope / HuggingFace 的 lmstudio-community 组织下载)。
适用场景
用户想"看图/OCR/描述图片",且希望本地推理(隐私、离线、免费)时使用。若模型未加载或服务未开,先按下方启动步骤操作。
前提检查与启动(按需)
- 检查服务是否在运行(端口 1234):
Test-NetConnection 127.0.0.1 -Port 1234 -WarningAction SilentlyContinue -InformationLevel Quiet - 未运行则启动并加载模型(LM Studio 的 lms CLI 默认在
$env:USERPROFILE\.lmstudio\bin\lms.exe,路径因安装方式而异,也可直接用 LM Studio 图形界面启动服务):
模型 id:& "$env:USERPROFILE\.lmstudio\bin\lms.exe" server start # 模型通常会自动加载;未加载时用: & "$env:USERPROFILE\.lmstudio\bin\lms.exe" load qwen3.5-9b-uncensored-hauhaucs-aggressive & "$env:USERPROFILE\.lmstudio\bin\lms.exe" status # 确认 Server: ON + 已加载模型qwen3-vl-8b-instruct(官方 Qwen3-VL-8B,含 mmproj-Qwen3-VL-8B-Instruct-F16.gguf 投影)或qwen3.5-9b-uncensored-hauhaucs-aggressive(含 mmproj-Qwen3.5-9B-Uncensored-HauhauCS-Aggressive-BF16.gguf 投影)。 注意:mmproj 投影文件一旦丢失,模型会退化为纯文本(报 "does not support image inputs"),需重新下载并lms unload --all后重载。 注意:LM Studio 设置里enableLocalService必须为 true(或用 lms server start 启动)。
调用
$env:PYTHONIOENCODING='utf-8'
python "<skill目录>\scripts\analyze.py" --image "图片路径.png" --prompt "用一句话描述这张图片的内容"
--image:本地图片路径(自动缩放至宽 ≤ 256px、JPEG 压缩后 base64 传入,规避大图 token 膨胀)--prompt:对图片的问题/指令(默认 "用一句话描述这张图片的内容")--max-tokens:默认 2048。该模型是 reasoning 模型,思维链会先消耗 token;脚本默认带--reasoning-effort minimal压缩思维链,一般不会卡住。若仍出现"只有 thinking、正式回答为空",调大--max-tokens或加--reasoning-effort none。--reasoning-effort:minimal(默认)|low|medium|high|none(none 时改用 enable_thinking=false 参数关思维链)。--model:默认qwen3-vl-8b-instruct;换旧社区版用--model qwen3.5-9b-uncensored-hauhaucs-aggressive。
输出
stdout 打印 JSON:model / reply(正式回答)/ reasoning(思维链,可忽略)/ usage。模型返回空 content 且 usage 里 reasoning_tokens 占满时,调大 --max-tokens 重试。
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 11d ago First seen · 58 lines · 57 tokens per session scan A 63eb010298ae
lino-local-vision is a skill published in the GitHub repository LINoone516/lino-agent-skills (4 stars, last pushed 1mo ago), licensed MIT. It adds 57 tokens to every session and 1,243 once invoked, about $0.0003 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 2026-08-31.
Other skills, from other repositories
vision
Call vision models (Doubao, Qwen, DeepSeek, OpenAI) to analyze images. Use when you need to understand screenshots, UI layouts, diagrams, or any image content. Supports png/jpg/webp/gif.
automodel
Guides the user in discovering a better model structure (e.g. from feature transformations in logistic regression to equation terms within PDEs to neural network layer compositions) out of data. Use when the user asks to create or improve an existing model. Uses an iterative meta/inner agent loop to explore structural…
umap-learn
Use UMAP-learn for nonlinear dimensionality reduction, 2D/3D embeddings, clustering preprocessing, supervised or semi-supervised UMAP, DensMAP, AlignedUMAP, and Parametric UMAP workflows.
bulk-rnaseq
End-to-end bulk RNA-seq orchestrator — takes raw FASTQ reads through QC and trimming (FastQC, fastp/Trim Galore), alignment and quantification (STAR, Salmon, featureCounts), assembles a gene-level counts matrix, then hands off to differential expression (pydeseq2), pathway/GSEA enrichment (pathway-enrichment), and…
shap
Explain and audit machine-learning predictions with SHAP. Use for selecting SHAP explainers and maskers, computing and validating feature attributions, handling multi-output explanations, and producing local or global SHAP visualizations.
genomic-intelligence
Predict regulatory features, gene structure, and expression directly from DNA sequence using Genomic Intelligence's hosted transformer DNA language models — no local GPU or model weights. Six tasks over a REST API and a hosted MCP server (keyless public demo): promoter regions, splice donor/acceptor sites, enhancer…