@tangle-network/agent-knowledge 22.0.9 → 22.0.11
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/CHANGELOG.md +8 -0
- package/package.json +3 -3
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,13 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## 22.0.11
|
|
4
|
+
|
|
5
|
+
The release check verifies each supported agent-eval minor at its newest published patch, not at x.y.0. agent-eval 0.217.0 was never published, so the publish of 22.0.10 failed while installing it. The package contents are those of 22.0.10, which admits agent-eval 0.218.
|
|
6
|
+
|
|
7
|
+
## 22.0.10
|
|
8
|
+
|
|
9
|
+
The agent-eval peer range admits 0.218 (`>=0.213.0 <0.219.0`). agent-eval 0.218.1 moved to agent-core `^0.13.1`, the version agent-runtime needs, so a consumer of agent-knowledge, agent-eval and agent-runtime installs one agent-core. Development runs on agent-eval 0.218.1.
|
|
10
|
+
|
|
3
11
|
## 22.0.9
|
|
4
12
|
|
|
5
13
|
A dead knowledge-improvement run no longer locks out its resume. An operator SIGKILLed inside `promoteKnowledgeCandidate` or `restoreKnowledgeCandidateBaseline` used to hold `run.lock.durable` for the 15-minute lease, so every re-run failed at once with `ELOCKED` until it expired; the run lease now records its holder the way the store lock does (#275), and the next run, promotion or restore takes it over at once when that record proves the holder dead, emitting an `AGENT_KNOWLEDGE_LOCK_TAKEOVER` warning. A re-run with the same candidate and the same activation then resumes the transaction and returns `recovered: true`. A live holder, or one that cannot be proven dead, still makes the next caller fail fast.
|
package/package.json
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@tangle-network/agent-knowledge",
|
|
3
|
-
"version": "22.0.
|
|
3
|
+
"version": "22.0.11",
|
|
4
4
|
"description": "Build, search, evaluate, and improve source-backed knowledge bases.",
|
|
5
5
|
"homepage": "https://github.com/tangle-network/agent-knowledge#readme",
|
|
6
6
|
"repository": {
|
|
@@ -85,14 +85,14 @@
|
|
|
85
85
|
"zod": "4.6.5"
|
|
86
86
|
},
|
|
87
87
|
"peerDependencies": {
|
|
88
|
-
"@tangle-network/agent-eval": ">=0.213.0 <0.
|
|
88
|
+
"@tangle-network/agent-eval": ">=0.213.0 <0.219.0",
|
|
89
89
|
"@tangle-network/agent-interface": "^5.2.0"
|
|
90
90
|
},
|
|
91
91
|
"devDependencies": {
|
|
92
92
|
"@arethetypeswrong/cli": "^0.18.5",
|
|
93
93
|
"@biomejs/biome": "^2.5.14",
|
|
94
94
|
"@neo4j-labs/agent-memory": "0.5.0",
|
|
95
|
-
"@tangle-network/agent-eval": "0.
|
|
95
|
+
"@tangle-network/agent-eval": "0.218.1",
|
|
96
96
|
"@tangle-network/agent-interface": "5.7.0",
|
|
97
97
|
"@types/node": "^26.6.2",
|
|
98
98
|
"mem0ai": "3.3.0",
|