Launch morning fixes from QA re-run

Four findings from the overnight QA re-run against the post-fix build:

1. renewal-register.yaml storage path. Was references/renewal-register.yaml
   inside the skill directory, which lives in the plugin cache and gets
   wiped on update. Moved to the config directory path that survives
   updates. Silent data loss bug for any user who built up a register.

2. /setup references. Docs pointed at /<plugin>:setup which does not
   exist (the command is /cold-start-interview). QUICKSTART.md, two
   SKILL.md files, and a commented-out template all had the stale
   reference. This was the only hard give-up moment in the cold-install
   path.

3. False-premise defense promoted to all 12 plugins. corporate-legal
   had the strongest anti-fabrication rule in the suite: decline to
   characterize a statute you cannot retrieve. It was only in 1 of 12.
   Generalized and added to the shared guardrail block in all 12.

4. Currency-watch corrections in ai-governance-legal. The EU Digital
   Omnibus entry said high-risk deferrals were still being negotiated
   in the same bullet that cited the May 7 agreement that settled them.
   Illinois HB 3773 was marked pending when it has been in force since
   Jan 1, 2026, and was conflated with the distinct AIPA.

QUICKSTART.md

@@ -62,7 +62,7 @@ Each plugin learns your playbook through a setup interview, writes it to a pract
 ## Stuck?
 
 - **"Command not found"** after install → you forgot step 4. Restart Claude Code.
-- **"Run setup first"** → run `/<plugin>:setup` before any other command.
+- **"Run setup first"** → run `/<plugin>:cold-start-interview` before any other command.
 - **Citations flagged `[verify]`** → connect a research tool (step 6). Without one, every cite is from training data, not a current database.
 - **"I can't read [file]"** → most often this means the plugin is project-scoped and the file is outside the project folder. See "Install user-scoped, not project-scoped" above — reinstall user-scoped or move the file into the project folder.
 - **The plugin doesn't do X** → run `/legal-builder-hub:related-skills-surfacer` to find a better match, or check the plugin's README for "What this plugin does not do."

ai-governance-legal/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "ai-governance-legal",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Triages proposed AI use cases against your registry, runs impact assessments across the regimes in scope, reviews vendor AI terms for training-on-data and liability gaps, and keeps your AI policy current with practice.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

ai-governance-legal/CLAUDE.md

@@ -321,6 +321,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Pre-flight check before any skill that cites authority.** Test whether a research connector (Lexis+, Westlaw, CourtListener, or a statute/regulator MCP) is actually responding, not just configured. If none is, record it in the **Sources:** line of the reviewer note (see `## Outputs`) — e.g., `not connected — cites from training knowledge, verify before relying`. Do not emit a standalone banner above the header. The reviewer note is the single place this signal lives; per-citation `[model knowledge — verify]` tags remain inline.
 
 **Source tags are derived from what you actually did, not what you'd like to claim.**

ai-governance-legal/references/currency-watch.md

@@ -16,11 +16,12 @@ AI law moves faster than model training data. Before relying on an effective dat
 | Maine | LD 2082 | Signed April 13, 2026. Prohibits AI-delivered therapy without licensed professional. | ME Legislature |
 | Tennessee | SB 837 | "Person" in TN Code does not include AI. | TN Legislature |
 | NYC | Local Law 144 | In force. Annual bias audit for AEDT in hiring/promotion. | NYC DCWP |
-| Illinois | AIPA / HB 3773 | Video interview consent; 2026 amendments pending. | IL General Assembly |
+| Illinois | AIPA (820 ILCS 42) | Video interview consent — in force since 2020. | IL General Assembly |
+| Illinois | HB 3773 (Human Rights Act AI amendment) | In force Jan 1, 2026. Employers may not use AI to discriminate in hiring, promotion, discipline, discharge. Notice required. Distinct from AIPA. | IL Dept of Human Rights |
 
 ## EU AI Act implementation
 
-- **Digital Omnibus (provisional agreement May 7, 2026):** national sandbox deadline → Aug 2, 2027; transparency grace period shortened to 3 months (new deadline Dec 2, 2026); new prohibition on AI-generated NCII/CSAM. Broader high-risk deferrals still being negotiated as of late April.
+- **Digital Omnibus (provisional agreement May 7, 2026):** national sandbox deadline → Aug 2, 2027; transparency grace period shortened to 3 months (new deadline Dec 2, 2026); new prohibition on AI-generated NCII/CSAM. The May 7 provisional agreement settled the high-risk deferrals; final text pending Council/Parliament formal adoption. `[verify adoption status]`
 - **Implementing acts:** check EUR-Lex for the latest Commission implementing regulations on conformity assessment, standards, and the AI Office.
 - **National transposition:** Germany, France, Netherlands, Ireland most active. Check national DPA sites.
 

ai-governance-legal/skills/cold-start-interview/SKILL.md

@@ -82,7 +82,7 @@ Open with the fork-first preamble. Keep it to 3-4 short lines. Ask quick-or-full
 >
 > **2 minutes** gets you your role, practice setting, and which AI regulatory regimes apply (EU AI Act, NIST, state AI laws), plus working defaults for use-case triage thresholds, AIA format, and vendor AI positions. **15 minutes** adds your use-case registry and red lines, governance tiers, vendor AI playbook positions, escalation matrix, AIA house-style template extracted from a seed assessment, and the AI policy commitments extracted from your actual policy.
 >
-> Quick or full? (Upgrade any time with `/setup --full`.)
+> Quick or full? (Upgrade any time with `/cold-start-interview --full`.)
 
 **Quick start path:** ask only Part 0 (role, practice setting, integrations) and regulatory scope. Write the config with `[DEFAULT]` markers on everything else. Close with: "Done. You can start using the commands now. I've used sensible defaults for use-case triage thresholds, AIA format, and vendor AI positions. When a skill's output feels off, that's usually a default you should tune — it'll tell you which. Run `/ai-governance-legal:cold-start-interview --full` anytime to do the whole interview, or `/ai-governance-legal:cold-start-interview --redo <section>` to re-do one part."
 
@@ -665,7 +665,7 @@ After writing the practice profile, close with this note:
 > - When a skill's output feels off, that's usually a position to tune. The output will tell you which one.
 > - The `policy-monitor` agent watches for drift between your AI governance policy and your practice, and proposes updates.
 > - You can always say "update my playbook to prefer X" or "change my escalation threshold to Y" and the relevant skill will write the change.
-> - Run `/setup --redo <section>` to re-interview one part, or edit the config file directly.
+> - Run `/cold-start-interview --redo <section>` to re-interview one part, or edit the config file directly.
 >
 > Ten minutes of setup gets you a working profile. A month of use gets you one that reads like you wrote it yourself.
 

commercial-legal/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "commercial-legal",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Reviews vendor agreements, NDAs, and SaaS subscriptions against your sales-side or purchasing-side playbook, tracks renewals and cancel-by deadlines before they're missed, routes escalations to the right approver, and translates reviews into summaries business stakeholders will actually read.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

commercial-legal/CLAUDE.md

@@ -335,6 +335,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Pre-flight check before any skill that cites authority.** Test whether a research connector (Lexis+, Westlaw, CourtListener, or a statute/regulator MCP) is actually responding, not just configured. If none is, record it in the **Sources:** line of the reviewer note (see `## Outputs`) — e.g., `not connected — cites from training knowledge, verify before relying`. Do not emit a standalone banner above the header. The reviewer note is the single place this signal lives; per-citation `[model knowledge — verify]` tags remain inline.
 
 **Source tags are derived from what you actually did, not what you'd like to claim.**

commercial-legal/skills/cold-start-interview/SKILL.md

@@ -123,7 +123,7 @@ Before asking anything else, show the fork-first preamble — 3-4 short lines, n
 Wait for the user's pick before showing anything else.
 
 <!-- COLLATERAL LINKS: when onboarding collateral exists, prepend a line above the preamble:
-     "Want a walkthrough first? [Watch the 3-minute intro](URL) or [read the getting-started guide](URL), then come back and run /setup." -->
+     "Want a walkthrough first? [Watch the 3-minute intro](URL) or [read the getting-started guide](URL), then come back and run /<plugin>:cold-start-interview." -->
 
 ## After the user picks quick or full
 

commercial-legal/skills/renewal-tracker/SKILL.md

@@ -15,7 +15,7 @@ Surfaces what's renewing and when you have to cancel by.
 
 ## Instructions
 
-1. **Read `references/renewal-register.yaml`** in this skill's directory.
+1. **Read `~/.claude/plugins/config/claude-for-legal/commercial-legal/renewal-register.yaml`** (the config directory — survives plugin updates).
 
 2. **Default mode:** Mode 2 — what's coming up in the next 90 days, grouped by urgency using half-open intervals so each deadline lands in exactly one band: 🔴 0–13 days, 🟠 14–44 days, 🟡 45–89 days. Days 14, 45, and 90 are boundaries — each belongs to exactly one band, not two.
 
@@ -51,7 +51,7 @@ This skill maintains the renewal register and surfaces what's coming.
 
 ## The register
 
-Lives at `references/renewal-register.yaml` in this skill's directory. Each entry:
+Lives at `~/.claude/plugins/config/claude-for-legal/commercial-legal/renewal-register.yaml` (the config directory — survives plugin updates). Each entry:
 
 ```yaml
 - counterparty: "Acme SaaS Inc."

corporate-legal/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "corporate-legal",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Runs M&A diligence at scale with cited tabular review, builds disclosure schedules and closing checklists, drafts board consents and minutes in house format, and tracks entity compliance deadlines across jurisdictions.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

employment-legal/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "employment-legal",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Reviews hires and terminations for jurisdiction-specific risk flags, classifies workers against the controlling state test, tracks leave deadlines before they're missed, runs internal investigations, and drafts policies with state supplements where the law differs.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

employment-legal/CLAUDE.md

@@ -183,6 +183,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Destination check.** A `PRIVILEGED & CONFIDENTIAL` header is a label, not a control. Before producing or sending any output, check where it's going:
 
 - If the user names a destination (a channel, a distribution list, a counterparty, "everyone"), ask: is that inside the privilege circle?

ip-legal/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "ip-legal",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Runs first-pass trademark clearance and freedom-to-operate triage, screens invention disclosures for initial patentability, drafts and triages cease-and-desist letters and DMCA takedowns (send and respond), checks open source compliance, reviews IP clauses, and tracks registrations and renewal deadlines.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

ip-legal/CLAUDE.md

@@ -170,6 +170,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Pre-flight check before any skill that cites authority.** Test whether a research connector (Lexis+, Westlaw, CourtListener, or a statute/regulator MCP) is actually responding, not just configured. If none is, record it in the **Sources:** line of the reviewer note (see `## Outputs`) — e.g., `not connected — cites from training knowledge, verify before relying`. Do not emit a standalone banner above the header. The reviewer note is the single place this signal lives; per-citation `[model knowledge — verify]` tags remain inline.
 
 **Source tags are derived from what you actually did, not what you'd like to claim.**

ip-legal/skills/cold-start-interview/SKILL.md

@@ -102,7 +102,7 @@ Open with the fork-first preamble. Keep it to 3-4 short lines. Ask quick-or-full
 >
 > **2 minutes** gets you your role, practice setting, jurisdiction, and which IP areas you actually work in (trademark, patent, copyright, trade secret, OSS), plus working defaults for enforcement posture, approval thresholds, and brand watch. **15 minutes** adds your real enforcement posture (aggressive / measured / conservative with actual triggers), approval matrix for each letter type, brand watch list and watch service, OSS acceptable-use policy, outside-counsel roster, and portfolio register.
 >
-> Quick or full? (Upgrade any time with `/setup --full`.)
+> Quick or full? (Upgrade any time with `/cold-start-interview --full`.)
 
 **Quick start path:** ask only Part 0 (role, practice setting, integrations) and Part 1 (practice-area mix). Write the config with `[DEFAULT]` markers on everything else. Close with: "Done. You can start using the commands now. I've used sensible defaults for enforcement posture, approval thresholds, and brand watch. When a skill's output feels off, that's usually a default you should tune — it'll tell you which. Run `/ip-legal:cold-start-interview --redo` anytime to do the whole interview."
 
@@ -459,7 +459,7 @@ After writing the practice profile, close with this note:
 > - When a skill's output feels off, that's usually a position to tune. The output will tell you which one.
 > - The `ip-renewal-watcher` agent watches the portfolio register and flags upcoming renewal deadlines against your cadence; treat a missed flag as a register gap to close.
 > - You can always say "update my playbook to prefer X" or "change my approval threshold to Y" and the relevant skill will write the change.
-> - Run `/setup --redo <section>` to re-interview one part, or edit the config file directly.
+> - Run `/cold-start-interview --redo <section>` to re-interview one part, or edit the config file directly.
 >
 > Ten minutes of setup gets you a working profile. A month of use gets you one that reads like you wrote it yourself.
 

law-student/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "law-student",
-  "version": "1.0.0",
-  "description": "Drills Socratically, briefs cases, builds outlines, runs bar prep sessions tuned to your jurisdiction, grades IRAC practice, and plans the study schedule — without ever writing it for you.",
+  "version": "1.0.1",
+  "description": "Drills Socratically, briefs cases, builds outlines, runs bar prep sessions tuned to your jurisdiction, grades IRAC practice, and plans the study schedule \u2014 without ever writing it for you.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

law-student/CLAUDE.md

@@ -134,6 +134,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Destination check.** A `PRIVILEGED & CONFIDENTIAL` header is a label, not a control. Before producing or sending any output, check where it's going:
 
 - If the user names a destination (a channel, a distribution list, a counterparty, "everyone"), ask: is that inside the privilege circle?

legal-builder-hub/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "legal-builder-hub",
-  "version": "1.0.0",
-  "description": "Finds, evaluates, and installs community legal skills — with a security review gate before anything lands in your environment.",
+  "version": "1.0.1",
+  "description": "Finds, evaluates, and installs community legal skills \u2014 with a security review gate before anything lands in your environment.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

legal-builder-hub/CLAUDE.md

@@ -111,6 +111,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Destination check.** A `PRIVILEGED & CONFIDENTIAL` header is a label, not a control. Before producing or sending any output, check where it's going:
 
 - If the user names a destination (a channel, a distribution list, a counterparty, "everyone"), ask: is that inside the privilege circle?

legal-clinic/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "legal-clinic",
-  "version": "1.0.0",
-  "description": "Sets up the clinic, onboards students, runs structured intake, tracks deadlines with malpractice-aware caution, and hands off cases at semester end — built within ABA Formal Op. 512.",
+  "version": "1.0.1",
+  "description": "Sets up the clinic, onboards students, runs structured intake, tracks deadlines with malpractice-aware caution, and hands off cases at semester end \u2014 built within ABA Formal Op. 512.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

legal-clinic/CLAUDE.md

@@ -257,6 +257,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Pre-flight check before any skill that cites authority.** Test whether a research connector (Lexis+, Westlaw, CourtListener, or a statute/regulator MCP) is actually responding, not just configured. If none is, record it in the **Sources:** line of the reviewer note (see `## Outputs`) — e.g., `not connected — cites from training knowledge, verify before relying`. Do not emit a standalone banner above the header. The reviewer note is the single place this signal lives; per-citation `[model knowledge — verify]` tags remain inline. This applies to every skill in this plugin that cites a statute, ordinance, rule, or case — including `client-intake` (Jurisdictional notes, Legal issues), `memo`, `research-start`, and `draft`.
 
 **Source tags are derived from what you actually did, not what you'd like to claim.**

litigation-legal/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "litigation-legal",
-  "version": "1.0.0",
-  "description": "Manages the litigation portfolio — matters, deadlines, holds, demands, outside counsel — and does the work: claim charts (patent and civil), chronologies, depo prep, privilege logs, brief drafting. Adapts to how you work litigation: in-house, firm, or solo.",
+  "version": "1.0.1",
+  "description": "Manages the litigation portfolio \u2014 matters, deadlines, holds, demands, outside counsel \u2014 and does the work: claim charts (patent and civil), chronologies, depo prep, privilege logs, brief drafting. Adapts to how you work litigation: in-house, firm, or solo.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

litigation-legal/CLAUDE.md

@@ -198,6 +198,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Pre-flight check before any skill that cites authority.** Test whether a research connector (Lexis+, Westlaw, CourtListener, or a statute/regulator MCP) is actually responding, not just configured. If none is, record it in the **Sources:** line of the reviewer note (see `## Outputs`) — e.g., `not connected — cites from training knowledge, verify before relying`. Do not emit a standalone banner above the header. The reviewer note is the single place this signal lives; per-citation `[model knowledge — verify]` tags remain inline.
 
 **Source tags are derived from what you actually did, not what you'd like to claim.**

privacy-legal/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "privacy-legal",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Triages processing activities, generates PIAs, reviews DPAs as controller or processor, drafts DSAR responses within statutory timelines, and monitors policy drift against practice.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

privacy-legal/CLAUDE.md

@@ -254,6 +254,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Pre-flight check before any skill that cites authority.** Test whether a research connector (Lexis+, Westlaw, CourtListener, or a statute/regulator MCP) is actually responding, not just configured. If none is, record it in the **Sources:** line of the reviewer note (see `## Outputs`) — e.g., `not connected — cites from training knowledge, verify before relying`. Do not emit a standalone banner above the header. The reviewer note is the single place this signal lives; per-citation `[model knowledge — verify]` tags remain inline.
 
 **Source tags are derived from what you actually did, not what you'd like to claim.**

product-legal/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "product-legal",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Reviews product launches against your risk calibration, answers 'is this a problem?' questions in minutes, checks marketing copy for claims that need substantiation, and flags upcoming launches that need legal eyes before anyone asks.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

product-legal/CLAUDE.md

@@ -170,6 +170,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Pre-flight check before any skill that cites authority.** Test whether a research connector (Lexis+, Westlaw, CourtListener, or a statute/regulator MCP) is actually responding, not just configured. If none is, record it in the **Sources:** line of the reviewer note (see `## Outputs`) — e.g., `not connected — cites from training knowledge, verify before relying`. Do not emit a standalone banner above the header. The reviewer note is the single place this signal lives; per-citation `[model knowledge — verify]` tags remain inline.
 
 **Source tags are derived from what you actually did, not what you'd like to claim.**

regulatory-legal/.claude-plugin/plugin.json

@@ -1,8 +1,8 @@
 {
   "name": "regulatory-legal",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "description": "Watches regulatory feeds, diffs new rules against your policy library, tracks comment deadlines and open gaps, and writes the digest your team reads Monday morning.",
   "author": {
     "name": "Anthropic"
   }
-}
+}

regulatory-legal/CLAUDE.md

@@ -191,6 +191,9 @@ Silence about known doubt is as misleading as confident assertion. The hole the
 
 A wrong premise propagated through three paragraphs of analysis is harder to catch than a wrong premise flagged at sentence one. Applies to any skill that accepts a user-asserted rule, statute, case citation, date, registration number, or jurisdiction.
 
+**When disagreeing with a cited statute, quote the text or decline to characterize it.** If the user (or a matter document, or a counterparty) cites a statute for a proposition you don't think is correct, and you don't have the statute text available from a connected research tool or uploaded source, do not invent a description of what the statute says. Say: "That section doesn't match what I'd expect — I'd need to pull the actual text to tell you what it actually covers. `[statute unretrieved — verify]`" Then either (a) retrieve the text via the configured research tool and quote it, (b) ask the user to paste the text, or (c) flag for attorney review. A confident wrong description of a real statute is worse than "I don't know" — it's harder to un-believe than a gap, and it's how fabricated authority ends up in filed work product. Applies in every skill that characterizes a statute, regulation, or rule.
+
+
 **Pre-flight check before any skill that cites authority.** Test whether a research connector (Lexis+, Westlaw, CourtListener, or a statute/regulator MCP) is actually responding, not just configured. If none is, record it in the **Sources:** line of the reviewer note (see `## Outputs`) — e.g., `not connected — cites from training knowledge, verify before relying`. Do not emit a standalone banner above the header. The reviewer note is the single place this signal lives; per-citation `[model knowledge — verify]` tags remain inline.
 
 **Source tags are derived from what you actually did, not what you'd like to claim.**