Improve README and add opt-in Jev workflows

.claude-plugin/skill-assets.sha256

@@ -3,4 +3,4 @@ c5d0c26f28c9ee14092f9deaf24c98dd8bef49d971fef2b7a537ffb1ab9f2887  .claude-plugin
 aeee7a94671ceb306fe2d24c5acc9f2d96ad8a8e7410536566799eea6265f080  skills/engraphis-memory/SKILL.md
 055655db84af07561d002f0c69744313d8413c39f3e873f941f0fa0b1e76dc66  skills/engraphis-memory/references/CONVENTIONS.md
 9d090a03f5b3f36a34d91f66b72c3844591f6915755ac3a6c6ba5f1b16977de5  skills/engraphis-memory/references/SCOPING.md
-68cd0add0d1025d495da2163e7c462b76b5fd1a4ff86b527fee49a4d8d6ab29b  skills/engraphis-memory/references/TOOLS.md
+db1db0df676e4c05fd80b878a1bd1ffc9f2045c0b0e01c0486b69ce0a5541b73  skills/engraphis-memory/references/TOOLS.md

.github/workflows/release.yml

@@ -1251,12 +1251,8 @@ jobs:
           notes_args=()
           latest_args=(--latest)
           promote_latest=true
-          if [ "$WAIVE_QUALIFICATION" = "true" ]; then
-            # A retained older repair must not displace a newer public Latest.
-            # The candidate comparison uses the existing public release history;
-            # failed lookups or unknown tag formats stop before any release write.
-            current_latest="$(gh release view --repo "$GH_REPO" --json tagName --jq .tagName)"
-            promote_latest="$(python - "$RELEASE_TAG" "$current_latest" <<'PY'
+          can_promote_latest() {
+            python - "$RELEASE_TAG" "$1" <<'PY'
           import re
           import sys
 
@@ -1267,7 +1263,13 @@ jobs:
 
           print("true" if version(sys.argv[1]) >= version(sys.argv[2]) else "false")
           PY
-          )"
+          }
+          if [ "$WAIVE_QUALIFICATION" = "true" ]; then
+            # A retained older repair must not displace a newer public Latest.
+            # The candidate comparison uses the existing public release history;
+            # failed lookups or unknown tag formats stop before any release write.
+            current_latest="$(gh release view --repo "$GH_REPO" --json tagName --jq .tagName)"
+            promote_latest="$(can_promote_latest "$current_latest")"
             latest_args=(--latest=false)
             if [ "$promote_latest" = "true" ]; then latest_args=(--latest); fi
             {
@@ -1293,7 +1295,14 @@ jobs:
               --repo "$GH_REPO" \
               --clobber
             if [ "$WAIVE_QUALIFICATION" = "true" ] && [ "$promote_latest" = "true" ]; then
-              gh release edit "$RELEASE_TAG" --repo "$GH_REPO" --latest
+              # Another publisher may have advanced Latest during asset upload.
+              # Recheck immediately before promotion; this narrows the race but
+              # does not provide an atomic lease across external publishers.
+              current_latest="$(gh release view --repo "$GH_REPO" --json tagName --jq .tagName)"
+              promote_latest="$(can_promote_latest "$current_latest")"
+              if [ "$promote_latest" = "true" ]; then
+                gh release edit "$RELEASE_TAG" --repo "$GH_REPO" --latest
+              fi
             fi
           else
             gh release create "$RELEASE_TAG" verified-dist/* release-evidence/* \

BENCHMARKS.md

@@ -93,15 +93,15 @@ interpretation and do not count as additional benchmark-quality gains.
 
 ### Public numeric evidence registry
 
-Every exact public aggregate retained below comes from the checked-in, public-safe
-[`offline-fixtures-v129.json`](docs/benchmark-evidence/offline-fixtures-v129.json) artifact. Its
-SHA-256 is
-`686394aa5d69cd296360defc5170d956c85565a5282c7766dc6466ba4c6aa6a5`, also recorded in the
+Every exact public aggregate retained below comes from the checked-in, public-safe
+[`offline-fixtures-pr252-four-gap-guards-20261007.json`](docs/benchmark-evidence/offline-fixtures-pr252-four-gap-guards-20261007.json) artifact. Its
+SHA-256 is
+`c53b82e4a847bf525c695c25a1813e8a591e86e39ca9e652e385e898ecd04010`, also recorded in the
 adjacent `.sha256` file. The artifact contains no raw questions, answers, prompts, customer data,
 or per-record content fingerprints.
 
-The fixture-suite digest is
-`8e2a562ba5d4c785c1e88481f9019446b43608109147cd7cd82597493004af35`. The artifact defines
+The fixture-suite digest is
+`e137eacdd70bbea7539dec6ed5cd4b59c980352a8e641094b449e474522aa109`. The artifact defines
 the digest algorithm and records the SHA-256 of every suite and dataset file. Each evidence ID
 also binds its exact command through `sha256(UTF-8 exact command)`:
 
@@ -121,12 +121,19 @@ means no number is claimed in this offline registry.
 The context-efficiency chart is generated from the registry values and the selected report schema.
 Historical LoCoMo, graph, handoff, consolidation, and security figures remain preserved in their
 source artifacts but are omitted from the current chart until each has a matching immutable,
-public-safe artifact. The chart labels coding outcomes, external datasets, and operational
-capacity as pending evaluation tracks rather than implying scores. Regenerate it with
-`python scripts/render_benchmark_report.py --report docs/benchmark-evidence/offline-fixtures-v129.json --output docs/images/context-efficiency.svg` after selecting the report to publish.
+public-safe artifact. The chart reports context size, candidate and packed retrieval quality, and
+grounded checks in separate panels; provider billing and MCP transport are not measured. Regenerate
+the SVG and matching PNG with:
+
+```bash
+python scripts/render_benchmark_report.py --report docs/benchmark-evidence/offline-fixtures-pr252-four-gap-guards-20261007.json --output docs/images/context-efficiency.svg --png-output docs/images/context-efficiency.png
+```
 
-The companion examples are also generated from that artifact with
-`python -m scripts.render_benchmark_examples --report docs/benchmark-evidence/offline-fixtures-v129.json --output docs/images/evidence-backed-agent-examples.svg`.
+The companion examples are also generated from that artifact with:
+
+```bash
+python -m scripts.render_benchmark_examples --report docs/benchmark-evidence/offline-fixtures-pr252-four-gap-guards-20261007.json --output docs/images/evidence-backed-agent-examples.svg --png-output docs/images/evidence-backed-agent-examples.png
+```
 The historical-to-executable mapping is in
 [`docs/BENCHMARK_CHANGE_COVERAGE.md`](docs/BENCHMARK_CHANGE_COVERAGE.md).
 
@@ -290,6 +297,27 @@ LoCoMo and LongMemEval retrieval diagnostics are retained as separate public-saf
 and checksum boundaries. Those values remain evidence-retrieval metrics, not end-to-end QA accuracy
 or an official LoCoMo leaderboard score.
 
+## Jev-assisted route-selection probe (provider-backed, exploratory)
+
+On 2026-10-03, the configured BYOK client ran `eval/jev_recall_quality.py` against 40 generated
+public synthetic tasks (10 each for lexical identifiers, graph relationships, temporal changes,
+and procedural context). Each call sent only the task query and at most two locally generated
+routes; no recalled memory bodies or stored user data were sent. Jev selected a route on all 40
+calls, but the baseline and Jev-assisted means were identical: nDCG@5 **0.740773**, Recall@5
+**0.75**, and answer-token coverage **0.75**. Every paired 95% interval was zero, so the
+improvement/non-inferiority gate did not pass. This synthetic probe does not establish a benefit
+on held-out user workloads, and no retrieval-quality improvement is claimed.
+
+The harness caps a run at 40 provider requests, makes at most one request per task, and requires
+explicit `--allow-remote` consent. Repeating it can consume BYOK provider usage:
+
+```bash
+python -m eval.jev_recall_quality --allow-remote --max-requests 40 --timeout 8
+```
+
+This exploratory provider run is separate from the registered offline artifact and the benchmark
+graphic below; it is not a general benchmark or customer-workload evaluation.
+
 ## What we do NOT yet claim
 
 - **No official end-to-end LLM QA accuracy.** The deterministic productivity agent measures the