











Every session teaches stndp how your code and team work — and it feeds that context back into your next prompt. The standup and the weekly board write themselves along the way.
Built for Claude Code ✦ supports other agents
stndp lives in the loop you already have.

✦✦✦A capture token from Account, five hook events in ~/.claude/settings.json. Nothing resident — a timer only retries failed sends.
Your conversation, the files you touched, the commits — read by the model, then gone. Only counts and SHAs are stored, never your code.




What holds up becomes memory, injected at session start and on every prompt. The classified summary lands on your weekly board.

Every session opens primed with three kinds of memory — your own conventions, your team's standards, and what last session shipped — learned automatically and kept honest by the audit.









Every session lands on the board as a summary you can actually skim — what shipped, what’s next, grouped by person and area. Always current, with no standup and no status-chasing.








A shared team project collects the whole crew’s sessions — reported by each member’s agent, grouped by person and area. Nobody writes a status update.















Each plan includes a monthly allowance. Use what you use.

Wire stndp into Claude Code in a few minutes. Every session after that teaches it — and writes itself up.