Entire

https://github.com/entireio/cli/pull/1347 check and validate these two review comments: cmd/entire/cli/checkpoint/checkpoint.go:428 — the CommittedInfo.CheckpointsCount comment: // CheckpointsCount is the total number of checkpoints across all sessions CheckpointsCount int This is the field surfaced by CommittedInfo (populated from summary.CheckpointsCount at committed.go:1270). After your change that aggregate is the sum of per-session prompt windows, not a count of checkpoints. The PR updated the sibling comments on WriteCommittedOptions.CheckpointsCount (line 230) and CommittedMetadata.CheckpointsCount (line 475), but this one on CommittedInfo was missed — so it now says exactly the thing the PR set out to stop claiming.

Doc gap (incomplete, not strictly wrong)

docs/architecture/sessions-and-checkpoints.md:214 — the CheckpointSummary JSON example shows "checkpoints_count": 3 with no prose explaining it, and doesn't mention the new save_step_count field at all. It's not actively wrong (no text claims it's a checkpoint count), but a reader cross-referencing the metadata schema won't learn that checkpoints_count is now a prompt count or that save_step_count exists.

3mo ago·5.1s

Validated against PR #1347.

CommittedInfo.CheckpointsCount comment — valid
- cmd/entire/cli/checkpoint/checkpoint.go:428 still says:
- But CommittedInfo.CheckpointsCount is populated from summary.CheckpointsCount in committed.go, and that summary is now the sum of per-session prompt-window/display “steps” counts.
- Suggested replacement:
Small nuance: I don’t see an updated comment directly on CommittedMetadata.CheckpointsCount; PR added SaveStepCount and its comment there, but CheckpointsCount itself remains uncommented.
Docs gap in sessions-and-checkpoints.md — valid
- docs/architecture/sessions-and-checkpoints.md:214 shows:
  with no prose explaining that this is now an aggregate prompt/display-steps count.
- save_step_count is not mentioned in that doc at all.
- The comment is right that this is incomplete, not strictly wrong.
- Since save_step_count is per-session CommittedMetadata, not root CheckpointSummary, it should probably be documented in prose or with a separate session-level metadata example, not added to the root CheckpointSummary JSON example.

No code behavior issue here; these are accurate comment/docs review points.