@windyroad/risk-scorer 0.3.6 → 0.4.0-preview.196

package/.claude-plugin/plugin.json

@@ -1,5 +1,5 @@
 {
   "name": "wr-risk-scorer",
-  "version": "0.3.6",
+  "version": "0.4.0",
   "description": "Pipeline risk scoring, commit/push/release gates for Claude Code"
 }

package/agents/pipeline.md

@@ -169,6 +169,47 @@ Include downstream back-pressure in the remediation list:
 
 Do NOT emit free-text "Your call:" or "consider splitting" prose. The structured `RISK_REMEDIATIONS:` block is the only output for above-appetite guidance.
 
+## Risk Register Hand-Off (Passive Trigger)
+
+When a pipeline run identifies a **register-worthy risk shape**, emit a structured `RISK_REGISTER_HINT:` block so the calling orchestrator can hand the finding off to `/wr-risk-scorer:create-risk` with pre-filled context. This is the passive trigger for the standing-risk register (`docs/risks/`) — it routes findings the pipeline already computes into the register without relying on the assistant remembering to invoke the create-risk skill (P110 / JTBD-001).
+
+### Trigger conditions (emit a hint when ANY fire)
+
+1. **Above-appetite residual** — any cumulative residual score exceeds appetite (> 4 per `RISK-POLICY.md`). A risk that breaches appetite on this change is a standing-risk candidate, not just a one-off remediation target.
+2. **Confidentiality disclosure** — the Confidential Information Disclosure check (below) flagged business metrics (revenue, user counts, pricing, client names, traffic volumes) in the diff. Confidentiality leaks are standing-risk-shaped even after the immediate remediation.
+3. **User-stated precondition** — the User-Stated Preconditions Check (below) flagged an unmet paired capability as a standalone Risk item. Unmet preconditions are standing-risk-shaped because the dependency gap persists until the paired capability ships.
+
+### Format (bulleted-list shape, multi-hint capable)
+
+A single pipeline run MAY surface more than one register-worthy shape (e.g. both an above-appetite residual AND a confidentiality leak). Emit one bullet per triggered condition:
+
+```
+RISK_REGISTER_HINT:
+- above-appetite-residual | <one-line prefill describing the risk>
+- confidentiality-disclosure | <one-line prefill citing what was flagged>
+- user-stated-precondition | <one-line prefill citing the unmet precondition>
+```
+
+### Reason-tag vocabulary (enumerated — reserved)
+
+The first column is one of exactly three reserved tags. Do NOT invent new tags; open new tickets to extend the vocabulary.
+
+| Tag | Meaning | Source section in this prompt |
+|---|---|---|
+| `above-appetite-residual` | Cumulative residual score > appetite | Above-Appetite Remediations |
+| `confidentiality-disclosure` | Business metric or client detail flagged in diff | Confidential Information Disclosure |
+| `user-stated-precondition` | Paired capability unmet; standalone Risk item | User-Stated Preconditions Check |
+
+The second column is free-form prose — a one-line prefill the orchestrator can pass to `/wr-risk-scorer:create-risk` as the initial risk title and description.
+
+### Consumption semantics (post-loop)
+
+The hint is consumed by the calling orchestrator **after** the ADR-042 auto-apply remediation loop converges (risk returns within appetite) or halts (Rule 5 exhaustion / user-aborted). Do NOT expect the hint to be consumed interleaved with remediation execution — if a remediation reduces the residual back within appetite, the above-appetite hint should still be recorded because the risk is standing even if this change is no longer in breach. The orchestrator decides whether to invoke `/wr-risk-scorer:create-risk` with the prefill or to defer to the user.
+
+### Silence guarantee (when no trigger fires)
+
+Do NOT emit `RISK_REGISTER_HINT:` when all cumulative scores are within appetite AND no confidentiality disclosure AND no user-stated-precondition fired. The hint is additive to the existing Below-Appetite Output Rule — a silent pass MUST remain silent. Do not emit an empty `RISK_REGISTER_HINT:` header with no bullets either — omit the block entirely.
+
 ## Confidential Information Disclosure
 
 Check diffs for business metrics (revenue, user counts, pricing, traffic volumes). Flag as a standalone risk if found.

package/agents/test/risk-scorer-register-hint.bats

@@ -0,0 +1,111 @@
+#!/usr/bin/env bats
+# Doc-lint guard: pipeline scorer MUST emit a structured RISK_REGISTER_HINT
+# block when a register-worthy risk shape is identified.
+#
+# Structural assertions — Permitted Exception to the source-grep ban (ADR-005 / P011).
+# These tests assert that the pipeline scorer's prompt defines the passive-trigger
+# contract required by P110. Agent prompts are specification documents; a
+# behavioural check of an LLM's output is out of scope for bats — the contract
+# document is what the PostToolUse hook and consuming orchestrator rely on.
+#
+# Background: P102 landed `/wr-risk-scorer:create-risk` as an on-demand
+# invocation surface for the risk register (docs/risks/). JTBD-001
+# (Enforce Governance Without Slowing Down) requires passive triggers that
+# fire "without a manual step" — assistant-remembering-to-invoke is the same
+# failure mode the job identifies as a pain point. P110 closes the gap by
+# having the pipeline scorer — which fires on every commit/push/release gate
+# — emit a structured hint line when it sees a register-worthy risk shape.
+# The calling orchestrator consumes the hint post-remediation-loop and
+# invokes /wr-risk-scorer:create-risk with pre-filled context.
+#
+# Cross-reference:
+#   P110:    docs/problems/110-risk-register-has-no-passive-trigger-slash-command-alone-partial-jtbd-001.open.md
+#   P102:    the parent ticket (slash-command MVP invocation surface)
+#   ADR-015: docs/decisions/015-on-demand-assessment-skills.proposed.md (Scorer Output Contract)
+#   ADR-013: docs/decisions/013-structured-user-interaction-for-governance-decisions.proposed.md
+#   ADR-042: docs/decisions/042-auto-apply-scorer-remediations-open-vocabulary.proposed.md (consumption timing)
+#   @jtbd JTBD-001 (enforce governance without slowing down — passive trigger)
+#   @jtbd JTBD-005 (invoke governance assessments on demand — composed via pre-filled create-risk)
+
+setup() {
+  AGENTS_DIR="$(cd "$(dirname "$BATS_TEST_FILENAME")/.." && pwd)"
+  PIPELINE="${AGENTS_DIR}/pipeline.md"
+}
+
+# ──────────────────────────────────────────────────────────────────────────────
+# Contract surface: RISK_REGISTER_HINT block exists and is documented
+# ──────────────────────────────────────────────────────────────────────────────
+
+@test "pipeline.md defines RISK_REGISTER_HINT block" {
+  # The structured hint line is the machine-readable passive-trigger contract.
+  run grep -q "RISK_REGISTER_HINT:" "$PIPELINE"
+  [ "$status" -eq 0 ]
+}
+
+@test "pipeline.md RISK_REGISTER_HINT uses bulleted-list shape" {
+  # Multi-hint capable: a single pipeline run can surface both an
+  # above-appetite residual AND a confidentiality-disclosure risk.
+  # Shape parallels RISK_REMEDIATIONS: (list-valued), not RISK_BYPASS_REASON: (single).
+  run grep -qE "^RISK_REGISTER_HINT:[[:space:]]*$" "$PIPELINE"
+  [ "$status" -eq 0 ]
+}
+
+# ──────────────────────────────────────────────────────────────────────────────
+# Trigger conditions: architect-approved three-trigger set
+# ──────────────────────────────────────────────────────────────────────────────
+
+@test "pipeline.md names above-appetite-residual trigger" {
+  # Trigger (a): any cumulative residual score > appetite.
+  run grep -q "above-appetite-residual" "$PIPELINE"
+  [ "$status" -eq 0 ]
+}
+
+@test "pipeline.md names confidentiality-disclosure trigger" {
+  # Trigger (b): confidential information (client names, revenue, pricing) flagged.
+  # Composes with the existing "Confidential Information Disclosure" section.
+  run grep -q "confidentiality-disclosure" "$PIPELINE"
+  [ "$status" -eq 0 ]
+}
+
+@test "pipeline.md names user-stated-precondition trigger" {
+  # Trigger (c): unmet user-stated precondition flagged as a Risk item.
+  # Composes with the existing "User-Stated Preconditions Check" section.
+  run grep -q "user-stated-precondition" "$PIPELINE"
+  [ "$status" -eq 0 ]
+}
+
+# ──────────────────────────────────────────────────────────────────────────────
+# Consumption semantics: hint is consumed post-loop, not interleaved
+# ──────────────────────────────────────────────────────────────────────────────
+
+@test "pipeline.md documents post-loop consumption semantics" {
+  # Architect advisory: the hint is consumed by the orchestrator AFTER the
+  # ADR-042 auto-apply remediation loop converges (or Rule 5 halts), not
+  # interleaved — otherwise register entries would be created for risks the
+  # loop might have remediated away.
+  run grep -qiE "post.?loop|after.*remediation.*(converge|loop)|after.*auto.?apply" "$PIPELINE"
+  [ "$status" -eq 0 ]
+}
+
+# ──────────────────────────────────────────────────────────────────────────────
+# Cross-reference: hint hands off to /wr-risk-scorer:create-risk
+# ──────────────────────────────────────────────────────────────────────────────
+
+@test "pipeline.md names the create-risk hand-off target" {
+  # The hint routes the consuming orchestrator to the MVP invocation surface
+  # from P102. Without the hand-off target, a hint line is undirected.
+  run grep -q "create-risk" "$PIPELINE"
+  [ "$status" -eq 0 ]
+}
+
+# ──────────────────────────────────────────────────────────────────────────────
+# Silence guarantee: below appetite + no special triggers = no hint
+# ──────────────────────────────────────────────────────────────────────────────
+
+@test "pipeline.md documents no-hint silence when within appetite and no triggers fire" {
+  # JTBD-001 "without slowing down": no hint noise when nothing is
+  # register-worthy. This parallels the Below-Appetite Output Rule for
+  # RISK_REMEDIATIONS: and protects the solo-developer 60-second review budget.
+  run grep -qE "[Dd]o NOT emit.*RISK_REGISTER_HINT|[Oo]mit.*RISK_REGISTER_HINT|no hint|no.*RISK_REGISTER_HINT" "$PIPELINE"
+  [ "$status" -eq 0 ]
+}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@windyroad/risk-scorer",
-  "version": "0.3.6",
+  "version": "0.4.0-preview.196",
   "description": "Pipeline risk scoring, commit/push gates, and secret leak detection",
   "bin": {
     "windyroad-risk-scorer": "./bin/install.mjs"