Summary
Worst-case gating depends on max_output / defaults (e.g. 1024). Uncapped calls MUTATE to a default; projection is coarse. Bad estimates → false HALTs or weak prevention.
Area: pre_call_worst_case, CallRequest.max_output_tokens, est_input
Desired behavior
- Better estimators: task/intent priors, model context window, historical p50/p95 for this agent+intent, optional caller hint.
- Keep fail-closed on unknown price; never price physical model max without a bound.
- Document accuracy limits; prefer MUTATE-to-priced-cap over crude HALT when headroom exists.
Acceptance sketch
- Measurable reduction in false-positive pre_call HALTs on demo workloads without raising overspend rate.
Reference: docs/notes/2026-07-23-design-fix-notes.md §13
Summary
Worst-case gating depends on
max_output/ defaults (e.g. 1024). Uncapped calls MUTATE to a default; projection is coarse. Bad estimates → false HALTs or weak prevention.Area:
pre_call_worst_case,CallRequest.max_output_tokens,est_inputDesired behavior
Acceptance sketch
Reference:
docs/notes/2026-07-23-design-fix-notes.md§13