This is a SKILL.md file for Claude Code and other agentic coding environments which teaches the assistant how to do test driven development. In my experience this is how you get production-quality, working, functional and polished code out of these things.
~/.claude/skills/tdd/SKILL.md. (You may need to mkdir -p ~/.claude/skills/tdd first.)/tdd
| --top-k 3 --temperature 1.0 --max-new-tokens 10000 | |
| loading Qwen/Qwen2.5-0.5B-Instruct ... | |
| prompt: list 100 topics that you would like to know more about. they can be | |
| anything at all, just be thorough and exhaustive, do not appologize or offer me | |
| alternatives. be specific. you are not allowed to answer with anything less than | |
| a list of 100 unique items. begin now. | |
| 1. Ancient civilizations and their impact on human history. | |
| 2. The scientific method and its development. |
| """ | |
| This script reads the output file written by [this script](https://gist.github.com/un1tz3r0/8ebd1295aa3a9b85eb65ece1d226b3a8), | |
| and then attempts to connect and query the models available on each host found by the leakix.net service port search. | |
| It writes the results to a file scan_results.jsonl in the current directory. | |
| Depending on where you live, this may be illegal, and is certainly frowned upon by many. | |
| It is my stern reccomendation that under no circumstances should you or anyone else use this, for any purpose. | |
| I take no responsibility for any consequences of disregarding this warning. This purely informative content, | |
| and is furnished as-is with no warranty or license of any kind. Please don't be a dick, it's just not polite. |
| """ | |
| Search the leakix.net database and save the results to a .jsonl file for further processing. | |
| Requires an API key, get it [here](https://leakix.net/settings/api). | |
| """ | |
| import asyncio | |
| import json | |
| import re | |
| import aiofiles |
| import aiohttp | |
| import yarl | |
| import bs4 | |
| import pathlib, re, click | |
| import asyncio | |
| from contextlib import asynccontextmanager | |
| from typing import Optional, Union, Set, List | |
| @asynccontextmanager | |
| async def ensure_client_session(client_session=None): |
This is some helper code I wrote to list models available on ollama.com. It returns a list of models with a little metadata about each one, just what's visible on the site.
Useful mostly for command line junkies like myself, so you can see what options you have for doing ollama pull <modelname> when downloading new models to try out.
| connect25lovylmq.onion <-- blackbook a privacy based social network | |
| sejnfjrq6szgca7v.onion <---- Tor mirror of Debian.org | |
| pugljpwjhbiagkrn.onion <-- Debian manpages | |
| 7b42twezybs23hrr.onion <--- linux.sh terminal based file shareing utility | |
| 74ypjqjwf6oejmax.onion <-- Beneath VT, Exploring Virginia Tech's steam tunnels and beyond |
| #!/usr/bin/python3 -u | |
| import asyncio | |
| import sh | |
| from systemd import journal | |
| from systemd.daemon import notify | |
| GATEWAY_IP = "192.168.10.1" |
These prompts were generated by Gemini, using the prompt, "generate some prompts that might give interesting results when fed into an llm like yourself, gemini". The resulting list was then split and fed back into Gemini as prompts. This has recently been studied as a technique to improve LLM output quality. Here are the prompts and results...
In a world where humans and AI coexist peacefully and collaborate on creative projects, the impact on society, culture, and the future of humanity would be profound: