meme-library

What meme-library does

meme-library is packaged as an Agent Skill so any skill-capable agent can drive it. It takes single images or batches, has a vision model tag each one, saves them under semantic filenames, groups them into series, and later selects the image that fits the current conversation. It only manages memes. Sending them is left to your own sender skill for WeChat, Telegram, Slack, Discord or another channel.

To keep the context small, a lightweight index.json holding each collection's vibe, best_for and avoid_when notes is always loaded, while a collection's meta.json is read only when the index points to it. search.py has a keyword-overlap fallback that tokenizes Chinese, and health_check.py reports orphan files, missing files, count drift and stale vibe fingerprints. An optional proactive mode asks the agent to judge on each reply whether a meme fits. A 15-image sample library is included. The README is mostly in Chinese, with an English section.

Key features

• Vision-based automatic tagging and semantic filenames for each image
• Series-based organization with a lightweight index loaded by the agent
• Context-based search with a keyword-overlap fallback that needs no embeddings
• health_check.py for orphan files, missing files, count drift and stale vibe notes
• Optional proactive mode with rules for when the agent should add a meme
• Sample library of 15 images in one collection

When to use it

• Having a Hermes chat bot pick a fitting sticker for a friend's good news
• Importing a downloaded sticker pack and letting the agent tag every image
• Checking a growing meme library for missing or stale entries

Who it is for: People who run a chat agent, especially Chinese-language chat bots, and want it to choose memes from their own collection.

How it fits with Hermes Agent

A general agent skill that names Hermes among its supported agents, and its installer checks for a Hermes setup first when detecting the platform.

How to install meme-library

These commands are copied from the project's README. Check the repository for the latest steps before you run them.

git clone https://github.com/yr-96/meme-library.git ~/.local/share/meme-library
bash ~/.local/share/meme-library/install.sh
python3 ~/.local/share/meme-library/scripts/verify_install.py

Requirements: A vision-model API key such as ANTHROPIC_API_KEY, which verify_install.py checks; a separate skill to actually send the chosen image

Note: The skill only manages and selects images; sending them requires a separate sender skill.

FAQ