Tetriz Engineering Kit
Open-source tools for everything around AI coding
10 followers
Open-source tools for everything around AI coding
10 followers
Agents write code. Help your team plan, understand, ship, and run it. 10 MIT-licensed tools for agent workflows, codebase learning, local CI, observability, tenant isolation, RPC access, log redaction, messaging, and GitHub attachments.
Free
Launch Team
AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted
Hey Product Hunter,
Your agents write code faster than your team can review and ship it.
So review becomes the bottleneck. And the code that does get merged isn't always code your engineers fully understand.
Today we're open-sourcing 10 engineering tools. It's how we work with agents ourselves.
Review that keeps up
- Agentic Workflow Kit brings fresh reviewers into Claude Code and Codex every few commits and carries each change through to a reviewed PR.
Understand what your agents shipped
- Codebase Trail turns your code and merged PRs into practice questions, so engineers understand what their agents shipped.
What AI-written code still has to get right
- 8 tools for local CI, Go linting, tenant isolation, log redaction, reliable messaging, observability and PR evidence.
If you tried Agent Desk, Session Intelligence or Tetriz Harness from our earlier launches: those show how your agents are set up and how their sessions go, and keep that setup current. These cover the work itself.
Where to start: Agentic Workflow Kit, if review is your bottleneck.
All 10 are MIT-licensed. Read the code, run it on a real branch, and tell us what breaks.
One question for teams building with coding agents: when do you bring in a reviewer? Every commit, every PR, or only when something looks off?
Code:
https://github.com/tetriz-ai/agentic-workflow-kit
https://github.com/tetriz-ai/codebase-trail
https://github.com/tetriz-ai/local-ci
https://github.com/tetriz-ai/wvet
https://github.com/tetriz-ai/observability-as-code
https://github.com/tetriz-ai/pgtenant
https://github.com/tetriz-ai/protoauthz
https://github.com/tetriz-ai/logredact
https://github.com/tetriz-ai/outbox-comms
https://github.com/tetriz-ai/ghattach
Built by the Tetriz team.
The review bottleneck framing rings true — I have found the hardest part is not generating code, it is trusting what got merged a week later. The Codebase Trail idea is the part I would reach for first: turning merged PRs into practice questions seems like a much better way to build real understanding than just re-reading diffs. I am curious how it handles a large, fast-moving repo — does it focus on recent changes, or try to cover the whole codebase at once?