A shared repository of battle-tested tools and primitives for AI coding agents — CI/CD wrappers, test runners, PR reviewers, deployment scripts, code quality checks, and other recurring patterns.
Why: Everyone keeps reinventing the same wheels. A PR reviewer agent written in isolation doesn't benefit from the lessons learned by the next person writing a similar tool. Shared, documented, reusable tooling would raise the collective baseline.
What it could look like:
- A monorepo (
agenticsnz/tools) with well-documented agent tools
- Each tool: a README explaining what it does, how to configure it, and known limitations
- Examples in multiple frameworks (Claude Code, Cursor, OpenCode, etc.)
- CI for the tools themselves — so they don't go stale
Categories to start with:
- PR reviewer / code reviewer
- Test generator / test runner
- Changelog / release notes generator
- Dependency updater
- Documentation generator
How to start: Survey the community (hackathon-style) to find what tools people have already built, then curate and document the best ones.
A shared repository of battle-tested tools and primitives for AI coding agents — CI/CD wrappers, test runners, PR reviewers, deployment scripts, code quality checks, and other recurring patterns.
Why: Everyone keeps reinventing the same wheels. A PR reviewer agent written in isolation doesn't benefit from the lessons learned by the next person writing a similar tool. Shared, documented, reusable tooling would raise the collective baseline.
What it could look like:
agenticsnz/tools) with well-documented agent toolsCategories to start with:
How to start: Survey the community (hackathon-style) to find what tools people have already built, then curate and document the best ones.