#ai-coding-agents

Git diff for AI agent repository instructions. Shows the blast radius of AGENTS.md and CLAUDE.md changes.

Usage and cost visibility for Claude Code and Codex — native macOS menubar app + cross-platform terminal CLI (npm, cargo, crates.io). ccusage-class reports, live 5h-block TUI, pure-Rust engine. Cost numbers are API-equivalent estimates.

Open-source server that refuses vulnerable package installs. The backend the refuse CLI shim calls.

Codex Issue Radar and maintainer-readiness tooling for AI coding agents.

Persistent repository conventions for Claude Code: rule files vendored into your repo, a drift and tamper checker, a branch guard hook, and skills to bootstrap, sync, and revise them.

Git-native shared context for teams whose members each use their own AI coding agent (Claude Code, Cursor, Antigravity, OpenCode...) on the same repo

Make agents prove their work. Independent verification, deterministic quality gates, and signed evidence for AI-generated code.

Run Claude Code, Codex, Copilot, Gemini, OpenCode and Amp side by side in isolated Git worktrees, with shared issue ownership, reservations and explicit handoffs.

Where did my Claude Code and Codex tokens go? One offline Python script: re-sent history, repeated tool output, files read twice.

Skill packs for Claude Code, Codex, and ~70 other coding agents: shipping, reviews, hardening, releases, product strategy, and go-to-market. MIT.

rustygrep: token-efficient grep for AI coding agents. ripgrep-class speed with --llm output compressed 60-95%, built-in MCP server for Claude Code, Cursor, OpenCode. Rust, SIMD-parallel, zero config. cargo install rustygrep

Fail-closed evidence gates for experiments and AI-agent workflows

職場 Shokuba — A Multi-Agent Harness. Watch, control, and verify a living team of AI software engineers in an isometric voxel office.

Agent-neutral Rust coding conventions + one-shot environment bootstrap (China-mirror aware). Works with Copilot, Claude Code, Codex, Cursor, Zed, Kimi Code, Gemini CLI.

Goal-based planning and proof gate for AI coding agents

A prompt for AI coding agents that generates a multi-tenant SaaS backend on AWS with two-layer tenant isolation - Lambda authorizer enforces tenant boundaries at the app layer, STS session policies enforce them at the AWS API layer. Terraform + Cognito + DynamoDB + Python 3.12.