Skip to main content
Get the session state, progress, and counts for bugs and agent verdicts. Add --full for the list of bugs, the sub-agents, and the latest message.

​
Arguments

​
string
required
The session to inspect.

​
Output

By default, the command prints:
completed, status, createdAt, lastActivityAt, and latestTurn can be absent on older servers; latestTurn is null before the first turn, and its times are null while unknown. progress shows whether Mo is making headway between messages:
  • testCases counts reported test cases by their latest verdict. A case with an active sub-agent counts as running; cases not yet started are the remainder.
  • suspectedBugs counts suspicions sent for reproduction that have no outcome yet.
  • agents.running counts sub-agents still working.
  • testingMinutes is the session’s testing clock. It runs only while Mo or a sub-agent works, and Mo can restart it once for each new message. timeBudgetMinutes is null when the session has no time limit.
Older servers omit progress and agents. --full adds:
blockerReason is present only when status is blocked or needs_you. With blocked it is access_blocked, login_required, or quota_exceeded; needs_you can also carry turn_failed. latestMessage is null when unavailable. kind is bug or flag (a static mistake recorded without reproduction). findings.bugs lists the main agent’s bugs and omits test cases and verdicts. agents lists each sub-agent with its agent kind, conversationId, status, task, and startedAt. Read one bug’s full details with qa bugs describe, or export everything with qa report. --output <file> writes the JSON to that file and prints the file’s path. status is the session’s state as the UI shows it: running, waiting_on_agents, checking (Mo’s last outcome is unknown), needs_you, blocked, ready (Mo marked the task complete), sleeping (Mo stopped without marking it complete), cancelled (someone stopped the session), or failed. Use it to decide what to do next; see what each status means. qa read and qa report report the identical status, so scripts can compare the commands directly. completed is the root agent’s explicit completion flag. It is not inferred from verdicts or a turn ending. Mo clears it when resuming work. Older servers omit it; an absent flag does not establish completion. Check status as well for active work, blockers, or input Mo needs.

​
Examples

List bug names:
Poll a session until it stops running, then list the bugs it found. A checking status that lasts a minute means Mo’s last outcome is unknown, so the loop stops:

​
Notes