Development log
Progress notes from building TaskCycle — releases, engineering decisions, and evidence.
- 2026-09-29
Different AI workflows. Shared engineering standards.
Concerns heard in workshops with software-house owners and CTOs: consistent quality, client-code privacy, and qualified review as coding agents become part of delivery.
- 2026-09-25
Why I built TaskCycle as an agentic development department
From structuring human software teams to building a governed team of AI agents: the origin story behind TaskCycle.
- 2026-09-21
How I keep long-running agentic tasks alive
Long-running agentic tasks are like intelligent batch processes - they should work without disturbing users until the task is done.
- 2026-09-11
What if your team could work all night to increase productivity?
Local AI agents can work around the clock, but the real question is whether a governed agentic team creates enough business value to justify its token and infrastructure costs.
- 2026-08-24
The Origin Story
Where this framework came from, and why the path to reliable autonomy is encoding acceptance criteria into tools the agent can be checked against — not hoping the agent infers what "correct" means.