LLMOps · Eval

A ruleset that treats every line of generated code as an admission of defeat.

Because verbose AI explanations waste tokens and patience, this skill makes agents talk like cavemen—terse, accurate, and cheap.

An AI job-search filter that scores listings, tailors your CV per role, and tracks the pipeline so you only apply to jobs worth keeping.

It strips the telltale signs of AI-generated writing from text, replacing statistical blandness with prose that sounds like it came from a person.

Headroom sits between your agents and the LLM to compress tool outputs, logs, and RAG chunks, delivering the same answers with a fraction of the tokens.

Context Mode is an MCP server that sandboxes bloated tool output and persists session memory in SQLite so AI agents don't lose their place when the context window compacts.

Hermes exists so you can stop re-teaching assistants that are supposed to remember, and run them on a $5 VPS instead of your laptop.

Ruflo turns Claude Code from a solo assistant into a coordinated swarm with shared memory, plugins, and cross-machine federation.

9Router is a local proxy that auto-switches your AI coding tools from paid to free providers and compresses token-heavy tool outputs so you stop hitting limits.

rtk intercepts common dev commands and feeds your AI only the compressed essentials, slashing token consumption by 60–90%.

Juggling Claude Code, Codex, Gemini CLI, and friends from separate terminals gets old fast.

A terminal tool that scores hundreds of local LLMs against your actual hardware so you stop downloading models that won't fit.

OpenHands exists so autonomous coding agents can plug into your workflow whether you want a Python library, a terminal tool, or a full cloud suite.

It gives LLM agents real brokerage access, then cages live trading inside strict mandates, kill switches, and audit ledgers.

Because stitching together LLM workflows, RAG, agents, and observability by hand is a full-time job.

It exists because AI context windows reset between conversations; Supermemory extracts facts, tracks changes, and recalls them later so agents don't start from scratch every time.

gstack turns Claude Code into a structured engineering team so a single founder can ship like a twenty-person startup without writing most of the code.

Because 'the model said so' is not an explanation, SHAP uses Shapley values from game theory to assign every feature its exact contribution to any prediction.

A living archive of extracted system prompts from Anthropic, OpenAI, Google, and xAI that exposes the hidden instructions shaping tone, tools, and guardrails.

Sub2API pools AI subscriptions behind a metered gateway so teams or resellers can distribute API quotas without building their own billing stack.