Skip to main content

pi-vision-helper

Use when the main model lacks vision capability (its input list in ~/.pi/agent/models-store.json has no "image") but the user asks to describe, analyze, or recognize the content of an image, photo, screenshot, or picture — including Windows paths (C:\...) and WSL paths (/mnt/c/...). Sends the image(s) to a configurable vision model (default: gpt-5.6-luna via pi-registry) and returns a detailed description (Chinese by default).

Jump to install

Source facts

Repository
yceachan/ea-pi-extensions
Last source activity
August 16, 2026 at 02:42
Detected SKILL.md language
Chinese
Stars
3
Forks
0

Install options

The review-first prompt is selected by default. You can switch to a direct command or download a local copy.

Review the source files

Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.