Institutional Knowledge
Your team
gets smarter.
Turn useful work into shared knowledge your team reviews before publishing, with optional memory you can choose to enable.
- How to fix widget timing bugtrace of 4 failed attempts → fixproposed
- Deployment checklist for servicesordered steps from successful deployproposed
- Auth troubleshooting flowchartpublished
Reviewed team memory
Facts your team learns and remembers. Salience-gated. Human-reviewed. Optional.
Salience-gated capture
Not every run gets remembered. High-value runs (you define the threshold) propose knowledge entries with ordered procedure traces and redacted evidence. Nothing auto-publishes.
Learning lane review
Proposed entries sit in /learnings for human review. Accept them to publish. Reject them to discard. Accepted drafts that cluster together become skill revision proposals.
Org + team scope
Memory is optional. Enabled via external memory-core container. User or team boundary: you choose. Disabled by default to keep every run fast.
SEE IT WORK
One agent. Real tasks.
UseAgentExample workflowOur widget service returns 500 on DELETE /widgets/:id when the widget has 3 or more linked components. Fix it.
Clone and search
Cloned repo, found DELETE handler in handlers/widgets.ts. Searched error logs. Every 500 has 'permission check failed' after component count check.
Reproduce locally
Set up test DB with 3 linked components, sent DELETE request, confirmed 500. Traced the stack: permission middleware runs after cascade check, causing the 500.
Fix the order
Moved permission check before cascade. Added test case: DELETE with 3+ components now returns 204. Old test suite still green.
Open PR
Created PR #487: 'Fix widget delete with linked components'. Added test case to prevent regression.
Questions, answered.
Is memory mandatory?
No. Memory is optional. Disabled by default. You enable it by configuring MEMORY_API_URL and the memory-core container.
Can agents write memory directly?
No. Memory is written only via the learning lane after the run completes, and only for high-value runs. Agents can read (recall) but not write.
Can one team member's memory be seen by another?
Depends on your scope. Memory can be user-scoped or team-scoped. Team scope means all team members see and add to the same memory. You choose when you enable it.
Start with one task.
See what you can hand off.
Explore a completed run, or join the early-access list to hear about getting your team started.