@DanielSmidstrup I’m building an SDLC engine that is harness agnostic, it currently supports OpenCode and PiDev, you can check it out here: https://t.co/eo11A8u4AS
@simonhamp For me, Git worktrees just became a time waste. I kept spending more time setting things up and switching contexts than actually working on the tickets. That’s why I built WorkDock to handle that part for me.
https://t.co/eo11A8u4AS
530M tokens, 36 tickets, and the team features taking shape. WorkDock is getting close to its first release.
Last week I announced WorkDock, an SDLC Orchestration Engine.
Since then:
→ 26 tickets done
→ 5 in progress
→ 1 under review
→ 1 blocked
Instead of focusing on managing AI harnesses, we can put our attention where engineering judgment matters:
→ Architecture.
→ Task definition.
→ Code review.
→ QA.
→ Product decisions.
Context switching doesn't have to lead to burnout when you're focused on the right things
If you're working with multiple AI agents, the hard part is coordinating everything around it.
Want to take a closer look?
🌐 WorkDock: https://t.co/TfsQ9YwuhD
⭐ Open source: https://t.co/eo11A8u4AS
Supervising 5+ AI agents is harder than it sounds.
At some point, you spend less time building and more time coordinating sessions.
That made me ask:
Why isn't working with AI agents more like managing software development?
Then it handles the rest:
→ Creates the PR
→ Review and provide feedback
→ Continue the work
It doesn't replace your tools.
It connects them.
The idea: bring proven software engineering practices to AI agents.
Ever since we launched, phishing emails have been nonstop. Just read the SolPhishHunter paper, 8,000+ phishing txs on Solana, ~$1.1M lost. We’ve seen the same email hit everyone on the team. One simple defense: share them openly so no one clicks first by error.