Skip to main content

​
Recall

cmd/crewship/cmd_recall.go hits the journal ?q= FTS5 endpoint and renders results in a snippet form optimised for question-answering (“how did we ship the auth fix?”). Different mental model from crewship journal, which is for filtered tailing/auditing — the two share the underlying endpoint but differ in defaults, presentation, and intent.

​
crewship recall <query>

The query is everything after the command name (joined with spaces). Server caps q= at 200 characters; longer queries fail fast with query too long: <n> chars (max 200).
FlagTypeDefaultEffect
--limit <n>int20Max matches to return.
--since <window>string(unset)1h, 24h, 7d, or RFC3339. Sent server-side.
--crew <slug-or-id>string(unset)Filter by crew. Resolved to ID.
--agent <id>string(unset)Filter by agent ID.

​
Examples

crewship recall "auth migration"
crewship recall "rate limit" --since 30d --limit 30
crewship recall "keeper denied" --crew backend-team
crewship recall "deploy" --format json | jq '.[].summary'
Each hit renders as a snippet card with timestamp, entry type, scope (crew/agent), the summary, and a body excerpt windowed around the first match. Matches are highlighted in bold ANSI when output is a TTY. 
2026-04-30 10:13  [peer.escalation]  backend-team / viktor
  agent escalated to lead — repeated DB lock timeout
  ...lock contention on credentials table during reload...

​
Common errors

  • query too long: <n> chars (max 200) — server-side limit.
  • bad --since: <reason> — value did not parse as duration or RFC3339.

​
See also