I recently made halftrace - point it at your OpenAI/Anthropic/LangSmith agent logs and it tells you what shape your agent's rule-following has, why, and what to try. I built it to try and find smooth decay over trajectory length on Claude.
I built WinLog because I kept seeing good engineers undersell themselves during reviews and promotions. Not because they weren’t doing great work, but because they simply forgot what they’d shipped over time.
WinLog helps you capture your "wins" as you go:
- You can log wins manually in a few seconds, or
- Passively capture candidates from Slack channels you choose (nothing is stored without approval)
Each win is rewritten into a concise, professional bullet you can use for:
- Performance reviews
- 1:1 prep
- Promotion packets
- Brag docs / weekly updates
It’s early, but I’ve been using it as my own tool and found it to be very helpful!
Would love feedback, especially from people who’ve struggled with review season or managing reports.
I recently launched https://winlog.dev, a Slack-first tool to help engineers capture their work as it happens so they’re better prepared for reviews and promotions.
It’s early days, but I’m using it with my own reports and so far it's been really helpful. The goal for 2026 is to grow this into a meaningful side-business, but if not, then it's useful to me!
I interviewed with Stripe early this year, and whilst sadly we couldn't go forward due to the fact that I'm British and was too late for an H1B - it was one of the most pleasant interviews I have ever had.
Contact - ruairidhwm[AT]gmail[DOT]com