@framers/agentos-skills-registry 0.5.0 → 0.6.0

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/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@framers/agentos-skills-registry",
3
- "version": "0.5.0",
3
+ "version": "0.6.0",
4
4
  "files": [
5
5
  "dist",
6
6
  "registry",
@@ -0,0 +1,37 @@
1
+ ---
2
+ name: topicality
3
+ version: '1.0.0'
4
+ description: Enforce allowed and forbidden conversation topics using semantic embedding similarity with session-aware drift detection
5
+ author: Frame.dev
6
+ namespace: wunderland
7
+ category: security
8
+ tags: [guardrails, topics, topicality, off-topic, embeddings, drift-detection]
9
+ requires_tools: [check_topic]
10
+ metadata:
11
+ agentos:
12
+ emoji: "\U0001F3AF"
13
+ ---
14
+
15
+ # Topicality
16
+
17
+ A guardrail automatically enforces conversation topic boundaries. Messages
18
+ matching forbidden topics are blocked. Messages outside allowed topics are
19
+ flagged. Gradual off-topic drift across multiple turns is detected.
20
+
21
+ ## When to Use check_topic
22
+
23
+ - To verify if RAG retrieval results are relevant to allowed topics
24
+ - Before presenting user-submitted content to the agent
25
+ - In content moderation workflows
26
+
27
+ ## What It Enforces
28
+
29
+ - **Allowed topics**: messages must be semantically related to at least one allowed topic
30
+ - **Forbidden topics**: messages matching a forbidden topic are blocked
31
+ - **Drift detection**: gradual off-topic steering across multiple turns is caught
32
+
33
+ ## Constraints
34
+
35
+ - Requires an embedding provider (OpenAI, etc.) to be configured
36
+ - Topic embeddings are computed lazily on first evaluation
37
+ - Drift detection tracks per-session state (cleaned up after 1 hour of inactivity)