mrxdev
Joined 9 months ago
Karma
5
Submissions
11
Member since
Jan 26
Recent submissions
Comment · 4 months ago
Comment · 7 months ago
Right on. Hope it’s of use! Feel free to open issues up on the repo if you bump into bugs.
Comment · 7 months ago
You hit the nail on the head. Completely agree.
This framing...
> Otherwise you can end up with persuasive agreement rather than real verification.
... is very familiar and one of the inspirational elements that led to https://github.com/spencermarx/open-code-review/tree/main (not trying to shamelessly plug, just your articulation was right on the money).
It's definitely the holistic set of static analysis and a robust automated test pyramid, then agentic analysis (ideally differentiated, and redundant with discourse -- like Open Code Review), that is proving to be effective at ensuring quality code.
That said, and a bit of a premature callout to the other project I mentioned, I think structured knowledge graphs to describe and commit to software intent (beyond what nested AGENT/CLAUDE/rule files and free form docs give you) are going to help us do a better job of engineering context and prompts for agentic development. This is the area that still feels like the wild west. Standardization is helping, and I think the toolset available now between RAG, lifecycle hooks, progressive disclosure in context and prompts, plus multi-agent implementation, review and discourse will help to close the loop and get us to the point of really confident and consistent pure AI-driven development. That's just my POV lol, I'll hop off my soap box now haha
Comment · 7 months ago
Great take. As it happens, I'm working with a number of other engineers on a solution that solves for this. However to dispel any false hope, static analysis + agentic development can only go so far. It's the holistic combo of these tools plus Agentic AI as both an implementation AND a feedback mechanism that unlock higher quality.
If you're looking to see what that feedback mechanism might look like in action, you might like checking out one of the other projects I've worked on which pre-dates Orc: https://github.com/spencermarx/open-code-review
Love where your head is at though! DEFINITELY an important problem we've got to solve here.
Comment · 9 months ago
Been working with spec-driven development since it first emerged (OpenSpec fan), and kept running into the same gap: specs get written, AI writes the code, but validating the implementation against the spec is still a manual slog.
So I put together a tool that runs multi-agent code reviews inside agentic AI environments (Claude Code, Cursor, Windsurf, etc).
The idea is simple – instead of one reviewer, you get a few different perspectives, modeled after common engineering teams: someone thinking architecturally, someone focused on code quality, another with security expertise, a QA engineer paying close attention to testing. They review changes independently, then discuss before synthesizing feedback. You can also configure custom reviewers and adjust redundancy if you want multiple sub-agents for a given role.
I originally built it to complement SDD tools like OpenSpec, but it drops into any codebase within an agentic environment (agentic IDE or agentic terminal).
My team has been running it against some larger codebases (~500k lines) and it's been genuinely useful – not as a replacement for human review, but as a way to catch things earlier. No doubt it has already saved us quite a bit of time.
```bash npx @open-code-review/cli init /ocr-review ```
Works as a pre-push sanity check or wired into CI.
Though warning: your token expenditure WILL increase, especially if you ramp up the number of reviewer types or their redundancy configs.
That said, curious what others think – is this useful? What's missing? Happy to hear feedback or ideas.
I've been there. That's why my team and I created Open Code Review (OCR), a free multi-agent code review tool that sits atop your existing agentic tools to give you high quality, structured code review designed to match code review on high performing engineering teams.
My teams and I have been dogfooding this for about a year now and we've seen as much as 70% time savings using OCR to ensure that human code review is only reached once code is truly polished and ready for real human eyes.
Save your code quality (and money) and try Open Code Review today.
Just dropped V2 with a number of stabilizations and enhancements. More to follow.
Link to get started here: https://github.com/spencermarx/open-code-review
Would love feedback, contributors, and really just to know that the tool is providing folks value.
Happy building