llm-pipeline

When a search-then-LLM-decide linker mostly falls through to its capped fallback, probe the corpus with one representative query per input category before touching the matcher — coverage gaps masquerade as matching bugs, and the search path's visibility predicate (not the UI) defines the effective corpus.

Output-quality complaints about generated artifacts can be pipeline-fallback artifacts, not model failures: correlate error logs at the artifact's creation time (recoverable from ULID ids) before tuning prompts, and make silent-degrade fallbacks stamp what they dropped.

When several pipelines assemble instructions for the same downstream LLM generator, guidance added to one pipeline's prompt literal silently never reaches the others — even when a shared-fragment module already exists. Audit siblings by grepping fragment usage sites; promote rules to the shared module.

Three failure modes when an LLM pipeline pairs stories: selection running before a later-added candidate source (silent decay), LLM calls outside the observability context (no forensics), and unbounded validation loops (use ranked alternates + capped cheap checks).