@windyroad/risk-scorer 0.14.0 → 0.14.1-preview.774

package/.claude-plugin/plugin.json

@@ -310,5 +310,5 @@
     }
   },
   "name": "wr-risk-scorer",
-  "version": "0.14.0"
+  "version": "0.14.1"
 }

package/agents/pipeline.md

@@ -442,4 +442,4 @@ They outrank the technical assessment.
 | 4 Significant | 4 | 8 | 12 | 16 | 20 |
 | 5 Severe | 5 | 10 | 15 | 20 | 25 |
 
-Label Bands: 1-2 Very Low, 3-4 Low, 5-9 Medium, 10-16 High, 17-25 Very High.
+Label Bands (ADR-086, supersedes ADR-065): 1-2 Very Low, 3-5 Low, 6-9 Medium, 10-16 High, 17-25 Very High. Only `score = 5` differs from the prior bands — now Low (was Medium), restoring feasibility for severe-but-rare residuals.

package/hooks/lib/risk-gate.sh

@@ -101,23 +101,26 @@ check_risk_gate() {
   fi
 
   # 5. Threshold check — block when the score EXCEEDS the project's
-  #    RISK-POLICY.md risk appetite (P007 / ADR-065). The threshold is the
-  #    adopter's documented appetite, not a code constant: a project whose
-  #    policy sets a higher appetite (e.g. "exceeds 9") must not have its
-  #    within-appetite changes gate-rejected.
+  #    RISK-POLICY.md risk appetite (P007 / ADR-086 supersedes ADR-065). The
+  #    threshold is the adopter's documented appetite, not a code constant: a
+  #    project whose policy sets a higher appetite (e.g. "exceeds 9") must not
+  #    have its within-appetite changes gate-rejected.
   #    Precedence: RISK_APPETITE env override > RISK-POLICY.md § Risk Appetite
-  #    parse > default 4. Default 4 reproduces the prior hardcoded `score >= 5`
-  #    behaviour exactly for integer scores (5 blocks, 4 passes) when the
-  #    policy is absent or unparseable. The parse is tolerant of the phrasings
-  #    "Threshold: N", "exceeds N", and "N/Low appetite", scoped to the
-  #    "## Risk Appetite" section. Cost ~3-8ms/invocation (ADR-065 § Consequences).
+  #    parse > default 5 (ADR-086). Default 5 tracks the new Low ceiling so
+  #    the no-policy adopter's default fallback admits residual=5 (the
+  #    Impact=5/Likelihood=1 floor for severe-but-rare risks). The normal
+  #    path is /wr-risk-scorer:update-policy interviewing the user to build
+  #    the policy; the default fallback is defensive.
+  #    The parse is tolerant of the phrasings "Threshold: N", "exceeds N",
+  #    and "N/Low appetite", scoped to the "## Risk Appetite" section.
+  #    Cost ~3-8ms/invocation (ADR-065 § Consequences carries forward).
   local DECISION
   DECISION=$(RISK_SCORE_VAL="$SCORE" RISK_APPETITE_ENV="${RISK_APPETITE:-}" python3 -c "
 import os, re, sys
 try:
     score = float(os.environ['RISK_SCORE_VAL'])
 except Exception:
-    print('no 4'); sys.exit(0)
+    print('no 5'); sys.exit(0)
 N = None
 override = os.environ.get('RISK_APPETITE_ENV', '').strip()
 if override.isdigit():
@@ -137,9 +140,9 @@ else:
             if m:
                 N = int(m.group(1)); break
 if N is None:
-    N = 4
+    N = 5
 print(('yes' if score > N else 'no') + ' ' + str(N))
-" 2>/dev/null || echo "no 4")
+" 2>/dev/null || echo "no 5")
   local DENIED="${DECISION%% *}"
   local APPETITE="${DECISION##* }"
 

package/hooks/test/risk-gate.bats

@@ -237,10 +237,12 @@ _write_matching_hash() {
 
 # ---------------------------------------------------------------------------
 # P007 / ADR-065 — the block threshold is the project's RISK-POLICY.md risk
-# appetite (block when score > N), not a hardcoded 5. Default N=4 when the
-# policy is absent or unparseable, which reproduces the prior `score >= 5`
-# behaviour exactly for integer scores. Precedence: RISK_APPETITE env >
-# RISK-POLICY.md parse > default 4. (Behavioural fixtures per ADR-052.)
+# appetite (block when score > N), not a hardcoded 5. Default N=5 per ADR-086
+# (was 4 under superseded ADR-065) when the policy is absent or unparseable,
+# which tracks the new Low ceiling so the default fallback admits residual=5
+# (the Impact=5/Likelihood=1 floor for severe-but-rare risks). Precedence:
+# RISK_APPETITE env > RISK-POLICY.md parse > default 5. (Behavioural fixtures
+# per ADR-052.)
 # ---------------------------------------------------------------------------
 
 @test "appetite 9 (exceeds 9): score 7 within the 5-9 band PASSES" {
@@ -275,29 +277,38 @@ _write_matching_hash() {
   assert_gate_denies "$TEST_SESSION" "commit" "appetite of 4/25"
 }
 
-@test "absent RISK-POLICY.md: defaults to appetite 4 (4 PASSES, 5 FAILS)" {
+@test "absent RISK-POLICY.md: defaults to appetite 5 per ADR-086 (5 PASSES, 6 FAILS)" {
   _use_policy ''   # no RISK-POLICY.md in the temp dir
   rm -f "$HASH_FILE"
-  printf '4' > "$SCORE_FILE"
-  assert_gate_allows "$TEST_SESSION" "commit"
   printf '5' > "$SCORE_FILE"
-  assert_gate_denies "$TEST_SESSION" "commit" "appetite of 4/25"
+  assert_gate_allows "$TEST_SESSION" "commit"
+  printf '6' > "$SCORE_FILE"
+  assert_gate_denies "$TEST_SESSION" "commit" "appetite of 5/25"
 }
 
-@test "unparseable RISK-POLICY.md (no appetite integer): defaults to appetite 4" {
+@test "unparseable RISK-POLICY.md (no appetite integer): defaults to appetite 5 per ADR-086" {
   _use_policy 'We are conservative about risk but state no number here.'
   rm -f "$HASH_FILE"
-  printf '4' > "$SCORE_FILE"
-  assert_gate_allows "$TEST_SESSION" "commit"
   printf '5' > "$SCORE_FILE"
-  assert_gate_denies "$TEST_SESSION" "commit" "appetite of 4/25"
+  assert_gate_allows "$TEST_SESSION" "commit"
+  printf '6' > "$SCORE_FILE"
+  assert_gate_denies "$TEST_SESSION" "commit" "appetite of 5/25"
 }
 
-@test "fractional score 4.5 FAILS under default appetite 4 (4.5 > 4)" {
+@test "fractional score 5.5 FAILS under default appetite 5 (5.5 > 5) per ADR-086" {
   _use_policy ''
   rm -f "$HASH_FILE"
-  printf '4.5' > "$SCORE_FILE"
-  assert_gate_denies "$TEST_SESSION" "commit" "4.5/25"
+  printf '5.5' > "$SCORE_FILE"
+  assert_gate_denies "$TEST_SESSION" "commit" "5.5/25"
+}
+
+@test "appetite via 'Threshold: 5' phrasing (ADR-086 new Low ceiling): score 5 PASSES, score 6 FAILS" {
+  _use_policy '**Threshold: 5 (Low)**'
+  rm -f "$HASH_FILE"
+  printf '5' > "$SCORE_FILE"
+  assert_gate_allows "$TEST_SESSION" "commit"
+  printf '6' > "$SCORE_FILE"
+  assert_gate_denies "$TEST_SESSION" "commit" "appetite of 5/25"
 }
 
 @test "RISK_APPETITE env override takes precedence over RISK-POLICY.md parse" {

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@windyroad/risk-scorer",
-  "version": "0.14.0",
+  "version": "0.14.1-preview.774",
   "description": "Pipeline risk scoring, commit/push gates, and secret leak detection",
   "bin": {
     "windyroad-risk-scorer": "./bin/install.mjs"

package/skills/create-risk/SKILL.md

@@ -113,7 +113,7 @@ Use the Risk Matrix from `RISK-POLICY.md`:
 
 - **Inherent Score** = Impact × Likelihood
 - **Residual Score** = Impact × Likelihood (after controls)
-- **Band** (for each) per the Label Bands table: 1-2 Very Low · 3-4 Low · 5-9 Medium · 10-16 High · 17-25 Very High
+- **Band** (for each) per the Label Bands table (ADR-086): 1-2 Very Low · 3-5 Low · 6-9 Medium · 10-16 High · 17-25 Very High
 - **Within appetite?** = residual score ≤ `RISK-POLICY.md`'s appetite threshold (read the threshold at runtime; do not hardcode)
 
 ### 5. Write the risk file

package/skills/update-policy/SKILL.md

@@ -108,20 +108,23 @@ Reference specific product features, user workflows, and infrastructure by name.
 - **Likely (4)**: High complexity, many code paths, or limited controls. Expected to occur without intervention.
 - **Almost certain (5)**: Known gap, no controls in place, or previously observed failure mode.
 
-**Risk matrix**: Include the Impact × Likelihood multiplication table (5×5 = scores 1-25) and the label bands:
+**Risk matrix**: Include the Impact × Likelihood multiplication table (5×5 = scores 1-25) and the label bands per ADR-086 (supersedes ADR-065):
 
 | Score Range | Label |
 |-------------|-------|
-| 1-4 | Low |
-| 5-9 | Medium |
-| 10-14 | High |
-| 15-25 | Critical |
+| 1-2 | Very Low |
+| 3-5 | Low |
+| 6-9 | Medium |
+| 10-16 | High |
+| 17-25 | Very High |
+
+The bands match the risk-scorer agent's authoritative source at `packages/risk-scorer/agents/pipeline.md` and the create-risk SKILL — single source of truth across the plugin. The Low ceiling at 5 admits residual=5 (Impact=5×Likelihood=1, the floor for severe-but-rare risks) within Low so an appetite of 5 is reachable for that class.
 
 The risk matrix is used by both the **risk-scorer agent** (pipeline risk assessment) and the **problem management process** (problem severity via `/problem` skill).
 
 ### 6. Confirm with the user
 
-You MUST use the AskUserQuestion tool (not plain text output) to collect user confirmation. Do not proceed to step 7 until you have received answers via AskUserQuestion.
+You MUST use the AskUserQuestion tool (not plain text output) to collect user confirmation. Do not proceed to step 6a or step 7 until you have received answers via AskUserQuestion.
 
 Call AskUserQuestion with a single message that presents:
 
@@ -129,6 +132,28 @@ Call AskUserQuestion with a single message that presents:
 2. The risk appetite threshold -- present the label bands from the agent contract (step 1) and recommend a threshold based on project maturity. Ask the user to confirm or adjust. A prototype with no real users may tolerate higher risk than a production system with paying users or compliance requirements
 3. Whether any business context is missing (e.g., compliance requirements, SLAs, user base size)
 
+### 6a. Tight-appetite warning when threshold < 5 (ADR-086)
+
+If the user picked an appetite threshold below 5 in step 6, fire a second `AskUserQuestion` confirm-with-warning before proceeding. The Low band's ceiling under ADR-086 is 5; an appetite below 5 means a class of risks (those with Impact=5/Severe and no impact-reducing control available) can never be within appetite — the policy is mathematically infeasible for that class. The user can still set the tighter threshold (some domains genuinely want to prohibit severe-impact activities), but the consequence must be a conscious choice, not a quiet trap.
+
+**Build the warning's example list** (cite concrete activity-classes the user is about to prohibit):
+
+1. Glob `docs/risks/R*.active.md`. For each file, read the `## Inherent Risk` and `## Residual Risk` sections and extract the numeric `Impact` value. Collect entries whose Inherent Impact = 5 OR Residual Impact = 5.
+2. If 1+ entries match: pick up to three. Format each as a kebab-case activity-class derived from the entry's filename slug (the part after `R<NNN>-` and before `.active.md`), followed by its R-ID as the audit pointer. Brief-before-ID per P350 — activity-class FIRST, R-ID in parentheses, never the R-ID alone as the carrier of meaning.
+3. If 0 entries match (empty register OR no Impact=5 entries): fall back to citing the policy's own Impact=5 row from the Impact Levels table just drafted — the user has it in working memory and it's the right surface for "what kind of thing would be prohibited."
+
+**The warning question**:
+
+> *"You've chosen appetite N. Under the rebalanced label bands (ADR-086) the Low ceiling is 5, so an appetite of N means residual-5 risks like `<activity-class>` ({R-ID}) and `<activity-class>` ({R-ID}) can never be within appetite under this policy — those activities become effectively prohibited (no amount of control work brings their residual below 5 because Impact=5 is fixed). Continue with appetite N, or revise?"*
+
+(Substitute the register-derived example list or the policy-row fallback as appropriate.)
+
+**Options**:
+- **Confirm: appetite N, with severe-impact activities prohibited** — the user explicitly accepts the trade-off. Proceed to step 7.
+- **Revise the appetite** — return to step 6 with the appetite question only.
+
+**Non-interactive (AFK) fallback** per ADR-013 Rule 6: if `AskUserQuestion` is unavailable, do NOT proceed — the warning is load-bearing (prohibits an entire risk class) and silently consuming it would normalise the prohibition. Halt with a clear "appetite < 5 selected interactively required" message for the orchestrator to drain later.
+
 ### 7. Validate draft with risk-scorer agent
 
 Before writing the policy file, invoke the risk-scorer agent to validate the draft. This is the gate -- the enforce hook will only allow writes to RISK-POLICY.md after the scorer returns PASS.