`todos show --json` omits several fields present in the raw API (completion_subscribers, due_on, starts_on, description_attachments, …)

[vmpn]

Summary

basecamp todos show <id> --json returns a data object with fewer fields than the underlying Basecamp 3 API endpoint (GET /buckets/{bucket}/todos/{id}.json), making the CLI output an incomplete view of a to-do.

CLI version

basecamp 0.7.2

Reproduction

ID=<some-todo-id>
PROJECT=<project-id>
BUCKET=<bucket-id>

# Fields surfaced by the CLI:
basecamp todos show "$ID" --in "$PROJECT" --json --jq '.data | keys'

# Fields the API actually returns:
basecamp api get "/buckets/$BUCKET/todos/$ID.json" --agent | jq 'keys'

Observed difference

Data-carrying fields present in the raw API but absent from todos show:

• completion_subscribers — the "When done, notify…" recipients shown in the web UI
• due_on — due date
• starts_on — start date
• description_attachments — file attachments embedded in the description
• comments_count — comment count on the to-do
• boosts_count — boost (reaction) count

A few hypermedia fields (completion_url, comments_url, boosts_url, subscription_url) are also absent — likely intentional, mentioning for completeness.

Impact

The missing fields aren't obviously absent — the call returns ok: true and a data object that looks complete. Automated callers (scripts, AI agents, downstream tooling) querying those fields via --jq get back null or [] with no way to distinguish "field omitted by the CLI" from "field genuinely empty on the server". For interactive use this is usually fine because the user can cross-check in the web UI, but for non-interactive callers it can lead to acting on incomplete information.

Workaround

Use basecamp api get "/buckets/<bucket>/todos/<id>.json" --agent and parse with jq directly.

---

Created with assistance of AI agent with human review.