TokenScope
Open-source token profiler for AI coding agents — see where your context budget goes
About TokenScope
TokenScope bills itself as the first open-source token profiler for AI coding agents. It analyzes session logs, detects repeated content and bloated tool results, and shows you exactly where your context window budget is being spent — with concrete fixes and estimates.
For anyone running Claude Code, Codex, or similar agents at scale, context bloat is a silent cost driver. TokenScope turns that invisible spend into a profile you can act on, trimming waste and keeping long agent sessions coherent and cheaper.
More Github
Spec Kit
Spec Kit is GitHub's official toolkit for Spec-Driven Development — a workflow where you write a clear specification first and let AI coding agents implement against it. It turns vague vibe-coding into a repeatable, reviewable process. The idea Rather than prompting your agent one feature at a time, you author a structured spec (a PRD-like artifact), and the agent plans, breaks down, and builds from it. That keeps large agent-built projects coherent instead of drifting. Spec-first workflow that scales beyond one-off prompts Works with Copilot and other coding agents Bridges product intent and generated code Get started on GitHub .
Superpowers
Superpowers is an open-source agentic skills framework and software development methodology built for AI coding agents. It packages battle-tested workflows — brainstorming, subagent-driven development, and structured SDLC practices — into reusable skills your agent can invoke on demand. Why builders care Instead of re-prompting the same instructions every session, Superpowers lets you codify how your agent should think and work. It leans into subagent-driven development, where a lead agent delegates focused tasks to specialized subagents, keeping context clean and output high-quality. Reusable skills that encode proven engineering workflows Subagent-driven development to parallelize and isolate context A methodology, not just a prompt pack — designed to compound over time Explore the framework on GitHub .
brain0
brain0 is a Rust-based "black box" for AI-written code. It builds a passive decision graph that links every commit back to the agent prompts behind it, giving you provenance, drift detection, and a DLP audit of exactly what your coding agents read and produced. It also ships MCP memory for coding agents plus signed, evidence-driven risk scoring. If you are shipping code written by Claude, Codex, or other agents, brain0 is a local-first way to answer "which prompt produced this change, and can I trust it?" — a governance layer builders increasingly need.
OpenScience
OpenScience is an open-source, agent-driven workbench built for scientific research. It packages a CLI, an LLM-backed "co-scientist" agent, and a set of research tools into a single TypeScript/Bun project so researchers and builders can run reproducible, AI-assisted investigations from the terminal. For AI builders, it is a strong reference architecture for domain-specific agents: it shows how to wire an LLM agent to real research workflows, structure tool use, and keep an experiment loop transparent and inspectable rather than a black box.
Godcoder
Godcoder is a local-first, open-source coding agent for your desktop built with Rust and Tauri. Bring your own LLM key; your code stays on your machine and only ever leaves to the model provider you choose. The agent builds and manages its own harness.