Skip to content

ruvnet.

Python

docs(ruvnet): finalize 2026-09-14 evidence

rUvd38753c88aaa
diff --git a/docs/self-improvement/2026-09-14-evidence.json b/docs/self-improvement/2026-09-14-evidence.json
index fe9c678..c5f85c1 100644
--- a/docs/self-improvement/2026-09-14-evidence.json
+++ b/docs/self-improvement/2026-09-14-evidence.json
@@ -3,7 +3,7 @@
"date": "2026-09-14",
"decision": "accept",
"baseline": "270fd513548e93862d59af8f22954c72686fb191",
- "candidateCommit": "3d8635856d62f1f0d0ca483fdc0a3e4899870937",
+ "candidateCommit": "d5f7171e79ee4ab93ddfbf43ba3b5f953d37341f",
"candidate": "bounded-upstream-snapshot-freshness",
"version": "0.2.2",
"upstream": {
@@ -18,7 +18,7 @@
},
"heldOut":{"baseline":{"passed":1,"total":6,"note":"The invalid-timestamp check threw only because the API was absent."},"candidate":{"passed":6,"total":6}},
"tests":{"groupsPassed":9,"groupsFailed":0},
- "validation":{"plugin":true,"skillsPassed":4,"skillsFailed":0,"marketplaceSchemas":2,"packageFiles":18,"compressedBytes":17560,"unpackedBytes":51600},
+ "validation":{"plugin":true,"skillsPassed":4,"skillsFailed":0,"marketplaceSchemas":2,"packageFiles":18,"compressedBytes":17636,"unpackedBytes":51834},
"benchmarks": {
"raw":"2026-09-14-benchmarks.json",
"pairs":15,
@@ -40,6 +40,6 @@
},
"memoryExperiment":{"catalogProjects":10,"changeRecords":6,"ruvectorWasm":"deferred","gnn":"rejected for this workload","persistentLearning":"not implemented","reason":"Corpus remains below the frozen 100-record minimum.","futureGate":{"minimumRecords":100,"minimumTaskSuccessGainPercentagePoints":10,"p95MaximumMs":20,"memoryMaximumBytes":134217728}},
"forecast":{"prediction":"Interfaces will increasingly surface freshness or invalidation state for cached evidence.","confidencePercent":80,"horizonDate":"2026-12-13","criterion":"At least two interoperable agent-memory protocols or major hosted connector interfaces expose freshness, expiry, or invalidation fields to clients.","sources":["https://arxiv.org/abs/2605.23986","https://arxiv.org/abs/2608.04746","https://blog.modelcontextprotocol.io/posts/2026-07-28/"]},
- "ci":{"run":"https://github.com/ruvnet/ruvnet/actions/runs/34833003257","status":"success","node22":"success","node24":"success"},
+ "ci":{"run":"https://github.com/ruvnet/ruvnet/actions/runs/34833554431","status":"success","node22":"success","node24":"success"},
"dashboard":"https://ruvnet-harness-evidence.ruv.chatgpt.site"
}
diff --git a/docs/self-improvement/2026-09-14.md b/docs/self-improvement/2026-09-14.md
index 40c3fce..2efd494 100644
--- a/docs/self-improvement/2026-09-14.md
+++ b/docs/self-improvement/2026-09-14.md
@@ -2,7 +2,7 @@
## Decision
-Accepted bounded upstream-snapshot freshness and published version 0.2.2. `ruvnet changes`, `ruvnet_changes`, and `ruv://changes` now expose `current`, `stale`, or `clock-skew`, an age, a 72-hour bound, and an expiry. Stale or future-dated evidence produces an explicit warning. The companion remains offline, read only, and incapable of refreshing itself.
+Accepted bounded upstream-snapshot freshness and published version 0.2.2. `ruvnet changes`, `ruvnet_changes`, and `ruv://changes` now expose `current`, `stale`, or `clock-skew`, an age, a 72-hour bound, and an expiry. The MCP resource recalculates freshness on every read, including in long-running hosts. Stale or future-dated evidence produces an explicit warning. The companion remains offline, read only, and incapable of refreshing itself.
## Upstream review
@@ -23,7 +23,7 @@ Fifteen alternating baseline/candidate pairs measured 10,000 discovery calls per
The deep Ruflo 3.25.6 code and dependency scans plus npm audit reported zero findings. Manual tracked-file credential patterns returned zero files. The secret scanner reported five files and STRIDE four, so coverage is limited and the result is not a claim of complete security. Residual risks include a fixed policy that may not fit every source, hosts treating MCP annotations as enforcement, and local system clocks affecting the age calculation.
-[Candidate commit](https://github.com/ruvnet/ruvnet/commit/3d8635856d62f1f0d0ca483fdc0a3e4899870937) passed [Node 22 and 24 CI](https://github.com/ruvnet/ruvnet/actions/runs/34833003257).
+[Final candidate commit](https://github.com/ruvnet/ruvnet/commit/d5f7171e79ee4ab93ddfbf43ba3b5f953d37341f) passed [Node 22 and 24 CI](https://github.com/ruvnet/ruvnet/actions/runs/34833554431).
## Memory, WASM and GNN decision