review-batch-failure-is-silent-skip
IN premise
When an LLM call fails for a review batch, the error is logged to stderr but the batch is skipped with no indication in the returned results; callers cannot distinguish "skipped due to error" from "no problems found."
Summary
If the LLM fails while reviewing a batch of beliefs, the system quietly drops that batch and moves on. Nothing in the output tells you it happened, so you might think everything was reviewed and passed when in reality some beliefs were never checked at all. This makes review results unreliable — a clean report could just mean the hard cases all errored out.
Dependents
These beliefs depend on this one:
- batch-fault-isolation-is-universal-across-llm-operations — Both LLM-facing batch operations — derive proposal application (try/except per proposal with error accumulation) and belief review (silent skip on per-batch LLM failure) — isolate faults at the individual item level, preventing any single bad item from aborting the entire batch.
- review-is-read-only-and-fault-tolerant — The review module operates entirely on in-memory snapshots with no storage dependency, handles missing antecedent references with placeholder text rather than exceptions, and silently skips failed LLM batches — achieving fault-tolerant read-only operation across all failure modes.
Details
| Source | entries/2026/05/05/reasons_lib-review.md |