@windyroad/risk-scorer 0.16.1-preview.900 → 0.16.2-preview.902

package/.claude-plugin/plugin.json

@@ -310,5 +310,5 @@
     }
   },
   "name": "wr-risk-scorer",
-  "version": "0.16.1"
+  "version": "0.16.2"
 }

package/hooks/lib/risk-gate.sh

@@ -38,7 +38,7 @@ check_risk_gate() {
   # 1. Score file must exist (fail-closed)
   if [ ! -f "$SCORE_FILE" ]; then
     RISK_GATE_CATEGORY="missing"
-    RISK_GATE_REASON="No ${ACTION} risk score found. Delegate to wr-risk-scorer:pipeline (subagent_type: 'wr-risk-scorer:pipeline') to assess cumulative pipeline risk."
+    RISK_GATE_REASON="No ${ACTION} risk score found. Delegate to wr-risk-scorer:pipeline (subagent_type: 'wr-risk-scorer:pipeline') to assess cumulative pipeline risk. Dispatch the scorer SYNCHRONOUSLY (run_in_background: false): a background-launched scorer does not fire its PostToolUse:Agent mark hook, so no marker persists and this gate re-blocks despite a within-appetite score (P402)."
     return 1
   fi
 

package/hooks/test/risk-gate.bats

@@ -60,6 +60,13 @@ assert_gate_allows() {
   assert_gate_denies "$TEST_SESSION" "commit" "No commit risk score found"
 }
 
+@test "missing-score deny instructs synchronous scorer dispatch (P402)" {
+  # A background-launched scorer's PostToolUse:Agent mark hook does not fire, so
+  # no marker persists and the gate re-blocks. The deny must tell the agent to
+  # dispatch synchronously. Mirrors external-comms-gate-canonical.bats.
+  assert_gate_denies "$TEST_SESSION" "commit" "run_in_background: false"
+}
+
 @test "score file with PENDING denies (non-numeric)" {
   printf 'PENDING' > "$SCORE_FILE"
   assert_gate_denies "$TEST_SESSION" "commit" "invalid value"

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@windyroad/risk-scorer",
-  "version": "0.16.1-preview.900",
+  "version": "0.16.2-preview.902",
   "description": "Pipeline risk scoring, commit/push gates, and secret leak detection",
   "bin": {
     "windyroad-risk-scorer": "./bin/install.mjs"

package/skills/assess-external-comms/SKILL.md

@@ -79,7 +79,7 @@ skill: wr-risk-scorer:external-comms
 prompt: <constructed review prompt from step 3>
 ```
 
-Wait for the wrapper to return. The wrapper invokes the external-comms agent internally; the agent's structured verdict block (`EXTERNAL_COMMS_RISK_VERDICT: PASS|FAIL` + optional `EXTERNAL_COMMS_RISK_REASON: ...` on FAIL) flows back verbatim. The `PostToolUse:Agent` hook (`risk-score-mark.sh`) fires on the wrapper's inner Agent invocation, derives the marker key from the prompt's `SURFACE:` + `<draft>` structure, and writes the marker automatically on PASS.
+Wait for the wrapper to return. The wrapper invokes the external-comms agent internally **synchronously** (`run_in_background: false`) — the `PostToolUse:Agent` mark hook fires reliably only for a synchronous agent, so a background-launched reviewer never persists its marker and the gate re-blocks despite a PASS verdict (P402 / P407). The agent's structured verdict block (`EXTERNAL_COMMS_RISK_VERDICT: PASS|FAIL` + optional `EXTERNAL_COMMS_RISK_REASON: ...` on FAIL) flows back verbatim. The `PostToolUse:Agent` hook (`risk-score-mark.sh`) fires on the wrapper's inner Agent invocation, derives the marker key from the prompt's `SURFACE:` + `<draft>` structure, and writes the marker automatically on PASS.
 
 **Do not write to `${TMPDIR:-/tmp}/claude-risk-*` yourself.** The hook is the only correct mechanism.
 

package/skills/external-comms/SKILL.md

@@ -23,10 +23,11 @@ This SKILL is an **invokable wrapper** around the `wr-risk-scorer:external-comms
 
 ### 1. Pass-through to the external-comms agent
 
-Invoke the external-comms subagent via the Agent tool with the caller's `$ARGUMENTS` verbatim. The `SURFACE:` line and `<draft>...</draft>` markers MUST be preserved exactly — the PostToolUse hook depends on the prompt structure for marker-key derivation:
+Invoke the external-comms subagent via the Agent tool with the caller's `$ARGUMENTS` verbatim. The `SURFACE:` line and `<draft>...</draft>` markers MUST be preserved exactly — the PostToolUse hook depends on the prompt structure for marker-key derivation. Dispatch it **synchronously** (`run_in_background: false`) — the `PostToolUse:Agent` mark hook (`risk-score-mark.sh`) fires reliably only for a synchronous agent; a background-launched reviewer's mark hook does not fire in time, so no marker persists and the external-comms gate re-blocks despite a PASS verdict (P402):
 
 ```
 subagent_type: wr-risk-scorer:external-comms
+run_in_background: false
 prompt: $ARGUMENTS
 ```
 

package/skills/pipeline/SKILL.md

@@ -20,10 +20,11 @@ This SKILL is an **invokable wrapper** around the `wr-risk-scorer:pipeline` agen
 
 ### 1. Pass-through to the pipeline agent
 
-Invoke the pipeline subagent via the Agent tool with the caller's `$ARGUMENTS` verbatim:
+Invoke the pipeline subagent via the Agent tool with the caller's `$ARGUMENTS` verbatim. Dispatch it **synchronously** (`run_in_background: false`) — the `PostToolUse:Agent` mark hook that writes the risk-gate marker fires reliably only for a synchronous agent; a background-launched scorer's mark hook does not fire in time, so no marker persists and the commit/push gate re-blocks despite a within-appetite score (P402):
 
 ```
 subagent_type: wr-risk-scorer:pipeline
+run_in_background: false
 prompt: $ARGUMENTS
 ```