-
-
Notifications
You must be signed in to change notification settings - Fork 205
Kanban and Task Progress
The Kanban view is a coordination surface across sessions and agents. It helps answer “what is moving, waiting, blocked, or done?” without reading every transcript.
Cards summarize normalized task state. They are most useful for selecting what to investigate next, not as the only source of truth. Open the underlying session when a state is surprising.
- Filter to the project or source you own.
- Scan active work for stale timestamps.
- Check waiting cards for unresolved user input or child-agent dependencies.
- Sample newly completed cards to confirm an actual result was produced.
- Open error cards and identify whether the failure is transient, recovered, or still blocking.
- Rapid transitions often reflect short agent/tool operations and are normal.
- Long active periods are reasonable only when activity continues or a long-running tool is expected.
- Waiting with active children usually indicates delegation, not a stall.
- Completed without an expected artifact deserves transcript review.
- Cards returning to active can happen when fresh imported or mirrored evidence corrects an older inferred state.
Agree on a shared meaning for “needs attention.” CCAM reports observed state; it does not own your delivery workflow. Avoid treating every waiting card as an incident or every completed card as reviewed.
For a handoff, share the stable session/task identifier and a link or filter description, not just a screenshot. Include the source name when multiple machines can use the same project path.
- Clear board filters.
- Refresh once and verify the live connection.
- Open the task's source session.
- Compare the latest event via CLI.
- If durable data is current, treat it as a client-state issue; if not, inspect ingestion.
See Reading the Dashboard and Troubleshooting Decision Tree.
Start Here · Localized product Wiki · Repository · Releases · Security · Report an issue
Guidance tracks the current master branch. For a deployed release, consult the repository documentation at that release tag before changing configuration.
CCAM Handbook
Use CCAM
- Read the dashboard
- Sessions and agents
- Kanban and tasks
- Tokens and costs
- Run agents
- Personal settings
Operate CCAM
- Deployment guide
- Hook health
- Security
- Backups and retention
- Prometheus and Grafana
- Remote machines
- Upgrades and rollback
- Incident runbooks
Automate and extend
- CLI patterns
- MCP workflows
- Alerts and webhooks
- Dynamic workflows
- Skills and plugins
- Safe config explorer
- Integration patterns
Help