History

cmd/crewship/cmd_history.go is the “what did I run lately?” command. Reads GET /api/v1/runs for the run list and, with --prompts, fetches GET /api/v1/chats/{chatId}/messages for the first user-role message of each run in a bounded 4-way worker pool. Different from crewship run list:

Adds an optional first-message preview per run.
Defaults to the last 24 h window.
Pretty single-line rendering with truncation.

`crewship history`

Flag	Type	Default	Effect
`--limit <n>`	int	`20`	Max runs to list. Must be > 0 (otherwise `bad --limit: must be greater than 0`).
`--since <window>`	string	`24h`	Time window — `1h`, `24h`, `7d`, or RFC3339. Applied client-side (the runs API has no `since=`). Inclusive on the boundary.
`--status <state>`	string	(unset)	Filter by status: `running`, `completed`, `failed`. Server-side `?status=`.
`--agent <slug-or-id>`	string	(unset)	Filter by agent. Resolved to ID before the request.
`--prompts`	bool	`false`	Also fetch the first user message per run. 4-way concurrent; failures degrade silently.
`--watch <duration>`	string	`""` (off)	Auto-refresh on an interval, e.g. `5s`, `1m` (min 1s). Shared `addWatchFlag` helper.

Examples

crewship history
crewship history --limit 50
crewship history --since 7d --status failed
crewship history --prompts                    # also fetch first user message
crewship history --agent viktor --format json | jq '.[].id'

Output:

2026-05-19 14:02  viktor             completed  CLI     "review the auth diff"
2026-05-19 13:51  petra              failed     ROUTINE
2026-05-19 13:38  default-agent      completed  CLI     "list open PRs"

Status colours: green (completed/succeeded), red (failed/error), yellow (running), gray (other).

See also

crewship retry <run-id> — rerun the same agent + recovered prompt.
crewship copy-prompt <run-id> — recover just the prompt text.
crewship recall <query> — search the journal by free text.
crewship explain <run-id> — summarize what a run did.