feat(turn): halt a turn stuck repeating one identical tool call

Ports grok-build's action stationarity breaker. A polling loop looks like
progress from inside the turn — every step succeeds — so nothing stops it
and it burns the whole budget re-reading one file. Nudge at 8 identical
batches, halt at 16, halt a pure no-op run at 4.

Three defects were found and fixed before this landed, none of which the
gates caught:

- The check runs AFTER execute_tool_calls, not before. Before it, the
  assistant's tool_calls are recorded with no results, so pushing the
  nudge triggers repair_dangling_tool_calls: synthetic "cancelled by the
  user" results get spliced in, the real results land after a user
  message, and dedup misses them — the orphan tool_result shape that
  draws a provider 400. It is Ok-gated too, since one `?` inside can
  leave a call resultless.
- `TurnOutcome::StationarityHalted` is its own variant grouped with
  Completed, not a reuse of Cancelled. As a cancellation it would report
  StopReason::Cancelled, fire the abort lifecycle, kill the turn's
  subagents, grow the goal harness back-off streak until three halts
  paused a running goal, and let completion-requirement recovery re-run
  the very turn halted for looping.
- The no-op gate is size-based, not name-based. kigi registers the shell
  tool as `run_terminal_cmd` and renames it `run_terminal_command`, so
  matching on "bash" silently disabled the tighter ceiling while every
  test still passed.
This commit is contained in:
ZacharyZhang-NY committed 2026-07-27 10:37:41 -04:00
1 parent a07d889ad9
commit ac23ebc9a1
8 files changed
+429 -5

No files matched your search

@@ -29,6 +29,13 @@ pub enum PromptCompletionKind {
MaxTurnsReached {
limit: usize,
},
/// One tool call repeated past its ceiling. Reported as `EndTurn`:
/// the turn ended, nobody interrupted it.
StationarityHalted {
tool_name: String,
run_len: u32,
true_noop: bool,
},
Rewound,
/// A queued prompt was removed (or cleared) from the server-authoritative
/// queue before it ever ran. Used to resolve the still-pending