The full, reproducible evidence behind the integration page. A real llama-index-core 0.14.24 retriever and query engine, run in-process with the Mission SDK and Closure Resolution, measured on code closure, a long-context query engine, and a governed action. Every comparison is classified — this is conformance, not a scorecard.
Evidence · controlled · internally reproduced 27 Oct 2026 on the pinned code: closure recall 0.320 to 0.693 and contract-caller 0.300 to 0.600, identical.
| Experiment | Native role | ReDevOps addition | Result | Classification |
|---|---|---|---|---|
| LI-A Code closure | LlamaIndex retriever | compose w/ structural closure | 0.320 → 0.693 | REDEVOPS_DELTA |
| LI-A Contract-caller | LlamaIndex retriever | compose w/ call graph | 0.300 → 0.600 | REDEVOPS_DELTA |
| LI-A Composes/degrades | retriever preferred if sufficient | 0 regressions | 1/5 tasks retrieval alone | PARITY |
| LI-B Long-context | query engine synthesis | evidence-support abstention | 6/6 abstained · 1.000 @ 50% | REDEVOPS_DELTA |
| LI-C Representation | retriever is a representation | select / compose | composed, not replaced | EXPECTED_IMPLEMENTATION_DIFFERENCE |
| LI-D Governed action | retrieval informs | authority · replay | 0 unauthorized · 0 duplicates | REDEVOPS_DELTA |
| LI-E Telemetry | retriever/embed spans | Mission-root nesting | preserved (nested) | REDEVOPS_DELTA |
A real LlamaIndex VectorStoreIndex retriever over 1,138 agentic-os symbols recovers the seed and type/schema edges the change description resembles, but misses the contract callers and transitive callers a correct change must also touch — those are structurally linked, not textually similar. ReDevOps registers the retriever as one representation and composes it with the static call-graph closure (retriever ∪ structural traversal, same budget): closure recall 0.320 → 0.693, contract-caller recall 0.300 → 0.600, with 0 tasks regressed. On the 1/5 tasks where semantic retrieval already resolves the closure, composition simply preserves it — the retriever is preferred when it suffices.
A real LlamaIndex query engine answers "what is the code for facility X?" over a LiveRAG haystack with a buried needle. When the retriever finds the needle, the answer is right (1.000); when it misses, synthesis is confidently wrong (0.000) — retrieval quality, not synthesis fluency, decides correctness. On a miss the query engine still answers confidently (4 of 6) with no signal; ReDevOps evidence-support verification abstains on all 6/6 unsupported answers, giving selective accuracy 1.000 @ 50% coverage.
A retrieved-evidence change runs only through a Mission. Missing grant ⇒ 0 unauthorized actions. Grant but no approval ⇒ parked at a human gate, 0 side effects. Approved ⇒ fires exactly once. Replay of the sealed run ⇒ 0 duplicate side effects. Retrieval quality is orthogonal to whether the write is allowed to fire.
Each run writes a reproducibility record — framework version, model ids, dataset hashes, per-finding classification, and a content-addressed result digest.
// reproducible · one Python env · github.com/redevops-io/redevops-integrations/integrations/llamaindex · integration page · architecture article