Install
$ agentstack add skill-yanqd0-my-claude-my-image-vision ✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.
Security review
✓ PassedNo issues found. Passed automated security review. · v0.1.0 How review works →
- ✓ Prompt-injection patterns
- ✓ Secret / credential exfiltration
- ✓ Dangerous shell & filesystem operations
- ✓ Untrusted network calls
- ✓ Known-malicious package signatures
What it can access
- ✓ Network access No
- ✓ Filesystem access No
- ✓ Shell / process execution No
- ✓ Environment & secrets No
- ✓ Dynamic code execution No
From automated source analysis of v0.1.0. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.
Verified badge
Passed review? Show it. Paste this badge into your README, it links to the public security report.
Reliability & compatibility
Declared compatibility
Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.
We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.
How agent discovery & health will work →About
对用户指定的图片进行预处理(可选),然后调用 DeepSeek V4 Vision API 返回文本描述。 厂商自动检测:当前配置为 DeepSeek 时走原生 Vision API,否则 fallback 到 Anthropic 原生路径。
执行步骤
- 解析意图:提取图片路径(相对/绝对路径均可)和预处理需求——用户可能说
"压一下再识别"、"灰度后看"、"把这张图模糊掉再 OCR"等。图片路径不存在时 用 AskUserQuestion(单选)确认,选项为邻近的候选图片路径。
- 确认预处理计划:列出拟执行的操作清单(0~N 个,需复合处理时标明合并为一张图),
用 AskUserQuestion(单选)确认——"按此预处理并识别"(推荐)、 "调整预处理(在 Other 描述)"、"不做预处理,直接识别"。
- 执行预处理(如需):
- 单图 1+ 操作:
~/.claude/skills/my-image-vision/scripts/preprocess.py
--op1 ... --opN -o
- 单图多处预处理后需合并 → 多次
preprocess.py后
~/.claude/skills/my-image-vision/scripts/composite.py ... --labels "标签1,标签2" -o 操作语法详见 Read references/preprocessing.md(条件触发,仅此步读)。
- 发送识图:
~/.claude/skills/my-image-vision/scripts/describe.py
[--prompt "指令"] [--max-tokens N]。 脚本自动从 ~/.claude/settings.json 读取 API key,默认走 DeepSeek Vision API。 按退出码分支处理:
- 退出码 0 → stdout 为识别文本,直接进入步骤 5;
- 退出码 23 → 当前配置非 DeepSeek,
Readreferences/anthropic-vision.md
按其中规范处理;
- 退出码 其他 → 错误,stdout/stderr 为错误信息,不尝试替代路径。
若需补充前置上下文(之前讨论的结构体定义、变量含义等),经 --prompt 传入 ——脚本为无状态单轮接口,skill 负责从对话中提取上下文组织提示词。
- 输出结果:将描述嵌入对话上下文。如涉及预处理操作链,简要标注(如
"(经灰度+二值化后识别)")。
约束
- API key 从
~/.claude/settings.json的env.ANTHROPIC_AUTH_TOKEN读取,
不传参数、不硬编码到脚本或 skill 正文。
- 图片仅送 DeepSeek Vision API(与当前对话模型同一供应商),不送第三方。
--prompt中不包含敏感信息(凭证、token 等)。
Source & license
This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.
- Author: yanqd0
- Source: yanqd0/my-claude
- License: MIT
Install and usage instructions live in the source repository linked above.
Reviews
No reviews yet, be the first.
Write a review
Versions
- v0.1.0 Imported from the upstream source.