Investigate AI-agent analysis report on the find_getmore_killcursors_commands specification. This ticket tracks findings identified during a systematic review of the spec: missing test coverage, ambiguous normative language, and spec/test inconsistencies.

Missing Tests

• find command is used instead of OP_QUERY when maxWireVersion >= 4
• Response includes cursor.id, cursor.ns, and cursor.firstBatch
• System collection queries use replacement commands (currentOp, fsyncUnlock, listIndexes, listCollections)
• Exhaust cursor: correct behavior for server 4.0 and earlier (OP_QUERY), 4.2–5.0 (OP_MSG preferred), 5.1+ (OP_MSG
  only)
• secondaryOk is set only on find, not on getMore or killCursors
• Negative limit/batchSize values are transformed to singleBatch=true and their positive equivalents
• On server 5.0+, cursor is closed when limit is reached and cursor ID > 0 is returned, requiring a killCursors
  call
• batchSize=1 returns a single document per batch without closing the cursor
• Tailable cursor: awaitData, maxTimeMS, and maxAwaitTimeMS handling
• getMore response: driver updates cursor ID, namespace, and stores nextBatch
• killCursors returns cursorsKilled, cursorsNotFound, cursorsAlive arrays
• Error handling: ok: 0 with errmsg and code fields

Ambiguities

• limit < 0 / batchSize < 0 transformation: Spec says to transform before adding to the find command but does
  not clarify at what point invalid combinations (e.g., limit=0 and batchSize=-1) should be rejected vs.
  transformed.
• maxAwaitTimeMS option requirement: "Drivers MUST provide a Cursor level option named maxAwaitTimeMS" — does
  "MUST provide" mean: accept the option, expose it in the API, or enforce a default value?
• Last step: getMore vs killCursors on limit exhaustion: Spec shows both as possible final steps but provides no
  algorithm for deciding which to issue.
• secondaryOk on getMore/killCursors (SHOULD NOT): If the server remembers the value from find, why is
  SHOULD NOT a requirement rather than a performance optimization?

Inconsistencies

• Changelog 2015-09-30 vs current spec: Old entry says "Legacy secondaryOk flag MUST be set to true on getMore
  and killCursors" but current spec says "SHOULD NOT." The contradiction is not resolved in the spec text. →
  DRIVERS-3482 /
  PR #1940
• Exhaust cursor 3.6 version gap: Changelog says "Exhaust cursors must use OP_MSG on 3.6+ servers" (2021-08-27) but
  current table says "4.0 and earlier MUST use OP_QUERY." Server 3.6 falls between these two rules — which applies?
• Duplicate maxTimeMS rules: One rule for non-tailable cursors and one for tailable with awaitData use nearly
  identical wording, creating confusion about which applies in which context.

Notes

• RST redirect at source/find_getmore_killcursors_commands.rst — can be cleaned up.
• Status: Accepted, Version 3.2. Stable spec.
• Extensive changelog from 2015–2024.
• FAQ section (lines 430–455) provides useful clarifications.