Adding Context
The fastest way is the context picker in the workspace input. It groups everything you can pull in:
Other sources attach from where they live:
Context Kinds
Under the hood each attachment is recorded against the run with a kind, so the agent receives structured material rather than a blob of text:
Small items are stored inline; large ones are saved once and linked by hash, so re-attaching the same file doesn’t duplicate it.
The Context Window Meter
Every attachment — and every turn of the conversation — spends tokens. The circular Context window indicator in the workspace shows how full the agent’s window is:
Hover it for the exact used / total token counts. Use it to decide when to start a fresh run instead of piling more context onto a saturated one.
Attaching context is additive and reviewable — chips show exactly what the agent will receive, and you can remove any of them before sending. Nothing is sent until your next message.
Related
- Diff Viewer — review and attach change sets
- Browser Panel — capture page elements as context
- Linked Tasks — pull issues from connected trackers
- Sentry — attach errors as signals
