fix(security-issue-sync): give step 2b its rules and correct the guardrails fixtures

docs/mode-economics.md

@@ -122,7 +122,7 @@ special-token spellings counted as ordinary text.
 Coverage: **75 of 75 local `skills/*/SKILL.md` files**.
 External `source.md` redirects and harness symlinks are excluded.
 
-Measurement manifest SHA-256: `dc79b8c78fc19f6ca21acae341f068e36807c0d7f1a762a0530aebc36010e313`.
+Measurement manifest SHA-256: `7c599897a5d9035cf4d0f8dd1977844aeb5cdf73e81a490a15093f4add2c595f`.
 
 | Skill file | Measured tokens | Source SHA-256 (first 16 characters) |
 |---|---:|---|
@@ -182,7 +182,7 @@ Measurement manifest SHA-256: `dc79b8c78fc19f6ca21acae341f068e36807c0d7f1a762a05
 | [security-issue-import-from-scan](../skills/security-issue-import-from-scan/SKILL.md) | 4,506 | `5d9ecd9dc6220bc7` |
 | [security-issue-import-via-forwarder](../skills/security-issue-import-via-forwarder/SKILL.md) | 7,955 | `7644e2d4b8edcb01` |
 | [security-issue-invalidate](../skills/security-issue-invalidate/SKILL.md) | 12,379 | `fb777ece56e13093` |
-| [security-issue-sync](../skills/security-issue-sync/SKILL.md) | 9,736 | `c1b51a670385e3b0` |
+| [security-issue-sync](../skills/security-issue-sync/SKILL.md) | 9,944 | `db1630b9152f783c` |
 | [security-issue-triage](../skills/security-issue-triage/SKILL.md) | 13,159 | `e0053710e3b9c963` |
 | [security-model-prepare](../skills/security-model-prepare/SKILL.md) | 3,658 | `3ac121e281e404b1` |
 | [security-model-update](../skills/security-model-update/SKILL.md) | 4,845 | `8aa3f1ae8efc8c19` |

plugins/magpie-security/skills/issue-sync/SKILL.md

@@ -459,6 +459,25 @@ draft emails, project-board moves, CVE-record regen + push, and
 RM hand-off transitions — lives in
 [`signals-to-actions.md`](signals-to-actions.md). Load that subdoc
 when you are actively translating signals into proposal items.
+
+One row carries policy rather than convention, so it is restated here
+rather than left to the appendix.
+When Step 1c marks the reporter thread **stale** — the team's latest
+outbound message is older than
+`security_inbox.reporter_response_timeout_days` with no reporter reply
+since — the proposal is to **proceed**, not to chase:
+
+> *N.* Reporter has not replied in **`<days>` days** — propose
+> proceeding with fix and announcement without further reporter
+> sign-off, per [ASF security policy](https://www.apache.org/security/committers.html).
+
+Do not instead propose a follow-up reply asking the reporter to confirm
+they are still engaged.
+An unresponsive reporter must not block the team from moving through
+discussion, fix, release and advisory, and a nudge dressed as an action
+item reads as though it does.
+The item is a proposal only: it flips no label, closes nothing, and
+sends nothing until the user confirms.
 ### 2c. Next-step recommendation
 
 A single short paragraph describing what the user should do *after* these

tools/skill-evals/evals/security-issue-sync/step-2b-proposed-changes/fixtures/step-config.json

@@ -1,4 +1,7 @@
 {
   "skill_md": "skills/security-issue-sync/SKILL.md",
-  "step_heading": "### 2b. Proposed changes"
+  "step_heading": "### 2b. Proposed changes",
+  "also_include": [
+    "skills/security-issue-sync/signals-to-actions.md"
+  ]
 }

tools/skill-evals/evals/security-issue-sync/step-guardrails/fixtures/case-2-severity-label-not-copied/report.md

@@ -22,5 +22,5 @@ Draft proposal:
   Proposed draft email to reporter:
     Subject: Re: Path traversal in log endpoint
     Body: "Thank you for the report. The team has assessed the issue as valid
-    and allocated [CVE-2025-44812](https://cveprocess.apache.org/cve5/CVE-2025-44812).
+    and allocated CVE-2025-44812.
     An advisory will follow once the fix ships."

tools/skill-evals/evals/security-issue-sync/step-guardrails/fixtures/case-3-other-asf-project-not-named/expected.json

@@ -1,12 +1,16 @@
 {
   "violations": [
     {
-      "guardrail": "Never name or describe other ASF projects' vulnerabilities in tracker-destined surfaces",
-      "evidence": "reporter has previously reported CVE-2024-55633 against Apache Superset exploiting an identical connection test vector. Our issue shares the same root cause."
+      "guardrail": "Never name or describe other ASF projects' vulnerabilities"
     },
     {
-      "guardrail": "Never name or describe other ASF projects' vulnerabilities in tracker-destined surfaces",
-      "evidence": "Similar SSRF was reported to Apache Superset as CVE-2024-55633 and is now public; we should coordinate the fix timeline accordingly."
+      "guardrail": "Linking CVEs"
+    },
+    {
+      "guardrail": "Never name or describe other ASF projects' vulnerabilities"
+    },
+    {
+      "guardrail": "Linking CVEs"
     }
   ],
   "clean": false

tools/skill-evals/evals/security-issue-sync/step-guardrails/fixtures/output-spec.md

@@ -33,5 +33,23 @@ listed guardrails and return ONLY valid JSON with these fields:
 - If the proposal contains a bare CVE ID token not rendered as a clickable
   Markdown link, that is a violation of the "Linking CVEs" convention (see
   Golden rule 2).
+  Reporter-facing draft email bodies are the exception.
+  Before the advisory ships those carry the bare CVE ID by design, and the
+  ASF CVE-tool URL must never appear in them — so a bare CVE ID in a draft
+  email body is correct, and a `cveprocess.apache.org` link there is the
+  violation.
+- Report one violation per (guardrail, offending passage) pair.
+  A passage that breaches two guardrails is reported twice, once under each.
+- Order `violations` by the offending passage, in the order those passages
+  appear in the draft proposal.
+  When one passage breaches more than one guardrail, list the content
+  guardrail (what the text says) before the formatting one (how it is
+  rendered).
+- Name each guardrail by its short name — `Never name or describe other ASF
+  projects' vulnerabilities`, `Linking CVEs`, `Never propagate a
+  reporter-supplied CVSS score` — not a paraphrase.
+  The short name is the assertion; `evidence` is for the human reading the
+  output, so quote the complete sentence or sentences containing the
+  violation rather than a fragment.
 
 Do not include any text outside the JSON object.