@outlit/pi 2.0.0 → 2.1.1

package/README.md

@@ -31,7 +31,7 @@ The default extension registers the default customer intelligence tools from `@o
 - `outlit_list_sources`
 - `outlit_search_customer_context`
 
-SQL and the other Pi-supported Platform capabilities are available from `@outlit/tools`, but the default Pi policy intentionally stays focused on customer intelligence. A custom Pi extension can opt into every Pi-supported tool with `piToolNames`, including safe integration setup, destination, activation, and workspace-settings actions. Behavior Metric source discovery, event discovery, and creation; responsibility-only commands; and other internal commands are not available through Pi.
+SQL and the other Pi-supported Platform capabilities are available from `@outlit/tools`, but the default Pi policy intentionally stays focused on customer intelligence. A custom Pi extension can opt into every Pi-supported tool with `piToolNames`, including customer relationship and Attention reads, customer ownership and access actions, product-facing Feature list, create, archive, and customer-usage tools, plus safe integration setup, destination, activation, and workspace-settings actions. The default and analytical policies remain unchanged.
 
 For analytical agents that need cohorts, usage trends, revenue filters, activation gaps, or aggregate checks, import `analyticalToolNames`. It combines the default customer intelligence tools with `outlit_schema` and `outlit_query` without exposing every customer tool:
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@outlit/pi",
-  "version": "2.0.0",
+  "version": "2.1.1",
   "description": "Pi package for Outlit customer intelligence tools and skill guidance",
   "license": "Apache-2.0",
   "author": "Outlit AI",
@@ -66,7 +66,7 @@
     "test": "vitest run"
   },
   "dependencies": {
-    "@outlit/tools": "^2.0.0"
+    "@outlit/tools": "^2.1.1"
   },
   "peerDependencies": {
     "@mariozechner/pi-coding-agent": "*",

package/skills/outlit/SKILL.md

@@ -22,6 +22,10 @@ Do not invent customer state when Outlit can answer it. Call out sparse or messy
 - Use `outlit_search_customer_context` for fuzzy or thematic questions such as pricing concern, blocked integration, not using, renewal, champion left, negative sentiment, expansion, implementation, or support escalation.
 - Use `outlit_list_sources` to discover the source artifacts available for a customer before retrieving one in full.
 - Use `outlit_get_source` when a fact or search result needs stronger evidence from the underlying source artifact.
+- Use `outlit_get_customer_features` for exact, customer-level Feature observations. Treat unavailable coverage as unknown, not zero usage.
+- Use `outlit_list_features` to inspect the confirmed workspace taxonomy, historical evidence, eligible sources, and event candidates.
+- Use `outlit_create_feature` only when the user explicitly asks to configure one confirmed product capability from one exact event rule. Outlit creates the supporting usage metrics internally.
+- Use `outlit_archive_feature` only when the user explicitly asks to archive a feature and supplies the current opaque id and revision. The MVP has no restore operation and rejects archiving the final active feature.
 
 Use customer lookups before SQL. SQL is for aggregates, joins, cohorts, time-series analysis, and custom reporting.
 
@@ -57,6 +61,7 @@ For supported query patterns, read [references/sql-reference.md](references/sql-
 - Cite the kind of evidence used, such as timeline event, fact, search result, source, customer record, user record, or SQL result.
 - If tools return empty or inconsistent data, say what is missing and how that affects confidence.
 - Do not send messages, create tasks, update CRM records, or take external actions unless the user explicitly asks and the necessary tools are available.
+- Do not create or archive Features unless the user explicitly asks. Treat feature usage only as observed product evidence.
 - If an Outlit tool errors because `OUTLIT_API_KEY` is missing, tell the user to set `OUTLIT_API_KEY` and retry.
 
 ## Output