Clarify recap scope and range controls

[peyton-alt]

https://entire.io/gh/entireio/cli/trails/335

Summary

• Show the actual recap window using the response since/until timestamps rendered in the local timezone.
• Show repo names in the summary context instead of only a repo count, with overflow text for larger multi-repo recaps.
• Add direct TUI range keys: d day, w week, m month, r 90d; keep t as a backward-compatible range cycle and move manual reload to R.
• Consume recap transcript availability diagnostics and render a short unavailable-transcripts note in the summary.
• Keep the diagnostics note wrapped inside the summary box at normal terminal widths.

Why

Users could see different recap numbers in different terminal windows because the CLI silently scoped to the current repo. The summary only said 1 repo, which made the scoped repo invisible. The range labels also did not explain the exact local-time boundaries behind day/week/month/90d.

The API can now report when checkpoint transcript payloads are failed, pending, or empty. Showing that count explains why checkpoint totals can be higher than mapped session totals without promoting missing transcript data into an Unknown agent row.

Testing

• mise run check
• ENTIRE_API_BASE_URL=https://partial.to go run ./cmd/entire recap --static --week --color never
• ENTIRE_API_BASE_URL=https://partial.to go run ./cmd/entire recap --week --color never

Smoke

• Against partial API version 72306b36-3ec2-49e2-82a9-c2d0910da6d6, the recap agent list shows Codex, Claude Code, and OpenCode; Unknown is not rendered.
• The summary still shows total checkpoints and includes 144 unavailable transcripts with the failed/pending/empty breakdown.

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Comment @cursor review or bugbot run to trigger another review on this PR

^{Reviewed by Cursor Bugbot for commit 0b4335f. Configure here.}

[Copilot]

Pull request overview

This PR improves the recap UX by making the recap “scope” (repo selection) and “range” controls clearer in both the static recap rendering and the TUI, including displaying an explicit server-provided time window when available.

Changes:

• Add explicit “window …” rendering for server-provided since/until, with configurable display timezone.
• Improve repo scope display by listing repo names (with an overflow “+N more”) when multiple repos are included.
• Update the recap TUI to support direct range hotkeys (d/w/m/r) and move reload to uppercase R, with corresponding tests.

Reviewed changes

Copilot reviewed 5 out of 5 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
cmd/entire/cli/recap/static_server_test.go	Updates/extends static recap tests for window display and multi-repo scope rendering.
cmd/entire/cli/recap/render_static.go	Adds window rendering, repo-scope text selection, and repo name normalization/deduping.
cmd/entire/cli/recap/me_recap.go	Extends API response model to include repos list.
cmd/entire/cli/recap_tui.go	Adds direct range keybindings and changes reload to R; refactors range setting into setRange.
cmd/entire/cli/recap_tui_test.go	Adds coverage for direct range keys and R reload; updates footer expectations.

Comments suppressed due to low confidence (1)

cmd/entire/cli/recap_tui.go:160

• The reload keybinding was changed to uppercase "R" (and lowercase "r" is now used for the 90d range), but the error screen still instructs users to "Press r to retry" (see View()). Pressing "r" will now change the range to 90d rather than retrying the current range. Update the error prompt and/or the handled key so the retry instruction matches actual behavior (e.g., use "R" for retry).

		case "R":
			m.requestID++
			m.loading = true
			m.loadErr = nil
			return m, m.fetch(m.requestID)