@cratis/pi 7.60.1 → 7.62.0

package/package/corpus/skills/cratis-screenplay-streams-and-consistency/references/consistency-and-concurrency.md

@@ -101,9 +101,11 @@ another event source) in `Provide()`/`Handle()` and returning a validation resul
 at commit is reported as a concurrency violation. Equivalently a Chronicle decision-consistency
 read with a `concurrency` scope on the event source or stream. A passive read is fresh
 computation, not an atomic read-and-append; a plain materialized view and a validator are
-advisory. Protection covers only a single projection keyed directly by the event source id (routing key
-empty or `$eventSourceId`). Reducers, joins, children, nested sources and **event-property
-routing** (fan-in from many streams by a payload reference) are refused, and a concurrency scope
+advisory under Arc read protection `Off` (the default); under `On` (Arc 22.106.0 or later) the read
+model parameters of `Handle()` and of a validator constructor are guarded. Protection covers a single
+projection keyed directly by the event source id (routing key empty or `$eventSourceId`) and an
+unfiltered reducer without constructor dependencies. Filtered reducers, reducers with dependencies,
+joins, children, nested sources and **event-property routing** (fan-in from many streams by a payload reference) are refused, and a concurrency scope
 on the referenced source does not see appends made on other sources. A rule over such a view, for
 example a course capacity computed from enrolment streams, needs option a or b: write the
 deciding fact on the source that owns the state, or give each scarce item its own identity. Sources: Arc `Documentation/backend/csharp/chronicle/read-models/injecting-into-commands.md` (limits unchanged);

package/package/corpus/skills/cratis-specifications-csharp/SKILL.md

@@ -14,8 +14,9 @@ focused assertions.
 
 Read [query and decision scenarios](../cratis-application-slice-specifications/references/query-and-decision-scenarios.md)
 when specifying query emissions, routed conflicts or protected projection reads.
-Call `UseDecisionReads()` before seeding actual events; pinned models cannot
-supply a guard. `Perform`/`Observe` need configured services/principal, and
+Seed actual events for a guarded command; pinned models cannot supply a guard.
+`new CommandScenario<T>()` selects the in-process decision log from the command's
+read-protection mode and dependencies, so `UseDecisionReads()` is not called. `Perform`/`Observe` need configured services/principal, and
 MongoDB observations still need real infrastructure.
 
 ## Verified product sources

package/package/corpus/skills/cratis-specifications-csharp/references/application-scenarios.md

@@ -16,7 +16,7 @@ call. A library or framework repository reaches for a scenario only to test the
 very engine it provides.
 
 Read [query and decision scenarios](../../cratis-application-slice-specifications/references/query-and-decision-scenarios.md)
-when using `QueryScenario<T>`, buffered emissions or `UseDecisionReads()`.
+when using `QueryScenario<T>`, buffered emissions or guarded command scenarios.
 Guarded decisions need actual seeded events, not pinned models; their given
 state and appends share one in-process log, unlike legacy dictionary seeding.
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@cratis/pi",
-  "version": "7.60.1",
+  "version": "7.62.0",
   "description": "Configuration-aware Cratis AI integration for Pi",
   "type": "module",
   "files": [