sint-ai · pshkv · May 30, 2026 · May 30, 2026
diff --git a/README.md b/README.md
@@ -740,6 +740,7 @@ docker-compose up
 - gRPC bridge guide: [`docs/guides/grpc-bridge-skeleton.md`](docs/guides/grpc-bridge-skeleton.md)
 - AutoGen interop fixtures guide: [`docs/guides/autogen-interop-fixtures.md`](docs/guides/autogen-interop-fixtures.md)
 - AgentSkill delegated authority fixtures guide: [`docs/guides/agentskill-authz-interop-fixtures.md`](docs/guides/agentskill-authz-interop-fixtures.md)
+- Industrial humanoid shipyard safety pack: [`docs/guides/industrial-humanoid-shipyard-safety-pack.md`](docs/guides/industrial-humanoid-shipyard-safety-pack.md)
 - action_ref identity/explainability profile: [`docs/specs/action-ref-identity-explainability-profile.md`](docs/specs/action-ref-identity-explainability-profile.md)
 - Payment governance profile (Economic Layer v1): [`docs/specs/payment-governance-profile-v1.md`](docs/specs/payment-governance-profile-v1.md)
 - Agent commerce governance profile: [`docs/specs/agent-commerce-governance-profile-v1.md`](docs/specs/agent-commerce-governance-profile-v1.md)

@@ -49,6 +49,7 @@ export default defineConfig({
           { text: "EU AI Act Conformity Pack", link: "/guides/eu-ai-act-conformity-pack" },
           { text: "Multi-Vendor Fleet Governance", link: "/guides/multivendor-fleet-governance" },
           { text: "Industrial Cell Safety Pack", link: "/guides/industrial-cell-safety-pack" },
+          { text: "Industrial Humanoid Shipyard Safety", link: "/guides/industrial-humanoid-shipyard-safety-pack" },
           { text: "Factory Action Pack Demo", link: "/guides/factory-action-pack-demo" },
           { text: "Regulated Consent Extensions", link: "/guides/regulated-consent-extensions" },
           { text: "Open-RMF Handoff Policy Receipts", link: "/guides/open-rmf-handoff-policy-receipts" },
@@ -84,6 +85,7 @@ export default defineConfig({
           { text: "Discord Launch Kit", link: "/community/discord-launch-kit" },
           { text: "External Contributor Onboarding", link: "/community/external-contributor-onboarding" },
           { text: "Independent Maintainer Scorecard", link: "/community/independent-maintainer-scorecard" },
+          { text: "Persona AI Shipyard Safety Brief", link: "/community/persona-ai-shipyard-safety-brief" },
           { text: "Robotics Collaboration Outreach", link: "/community/robotics-collaboration-outreach-schedule" },
           { text: "Sunnybotics Collaboration Brief", link: "/community/sunnybotics-collaboration-brief" },
           { text: "Sunnybotics Outreach Drafts", link: "/community/sunnybotics-outreach-drafts" },
@@ -112,6 +114,7 @@ export default defineConfig({
           { text: "Ecosystem Outreach 2026", link: "/roadmaps/ecosystem-outreach-2026" },
           { text: "Factory Action Pack Upgrade Sprints", link: "/roadmaps/factory-action-pack-upgrade-sprints" },
           { text: "Graph-First Coordination Upgrade", link: "/roadmaps/graph-first-coordination-upgrade-2026" },
+          { text: "Industrial Humanoid Shipyard Safety Sprint", link: "/roadmaps/industrial-humanoid-shipyard-safety-sprint" },
           { text: "Humanoid Robotics Integrations", link: "/roadmaps/humanoid-robotics-integrations-2026" },
           { text: "Post-Quantum Crypto Agility", link: "/roadmaps/post-quantum-crypto-agility" },
           { text: "Physical AI Governance", link: "/roadmaps/PHYSICAL_AI_GOVERNANCE_2026-2029" },

@@ -0,0 +1,117 @@
+# Persona AI Shipyard Safety Brief
+
+Status: outreach-ready technical brief
+
+Persona AI is building industrial humanoid robots for heavy industry, including
+shipyards, energy, construction, and manufacturing. Public materials emphasize
+work such as welding, fabrication, inspection, supervisor coordination,
+bystander awareness, and robot-to-robot collaboration.
+
+SINT can help by acting as the runtime governance and evidence layer between
+Persona-style humanoid intent and physical execution.
+
+## Core Fit
+
+Persona's problem is not only making a humanoid move or weld. In production,
+shipyards need to answer:
+
+1. Was this robot authorized to perform this action?
+2. Were the right permits active?
+3. Was the workspace safe for humans and other robots?
+4. Was the welding or manipulation path reviewed before execution?
+5. Can a supervisor, customer, or surveyor reconstruct what happened later?
+
+SINT is built for that boundary.
+
+## SINT Value for Persona-Style Deployments
+
+### 1. Shipyard Safety Policy Gateway
+
+Every high-consequence robot action can route through
+`PolicyGateway.intercept()` before execution:
+
+- start welding arc
+- start grinder
+- enter confined space
+- lift material
+- move through shared work zone
+- release robot-to-robot handoff
+- execute emergency stop rollback
+
+### 2. Hot-Work and Welding Evidence
+
+SINT can bind a welding request to:
+
+- hot-work permit
+- fire-watch readiness
+- fume extraction status
+- gas-monitor status
+- weld procedure reference
+- simulation receipt digest
+- offline program hash
+- supervisor approval
+
+### 3. ABS-Style Remote Survey Support
+
+SINT can export hash-chained evidence for:
+
+- policy decision and assigned tier
+- safety-controller context
+- robot and work-order identifiers
+- sensor calibration references
+- simulation receipt and program digest
+- event hash and previous hash
+
+This supports review and incident reconstruction. It does not replace
+classification society assessment or product approval.
+
+### 4. RaaS Fleet Governance
+
+For robot-as-a-service deployments, SINT can enforce customer/site policy:
+
+- per-site action limits
+- operator approval rules
+- tenant-specific permit requirements
+- update and model-change gates
+- evidence retention and export
+
+## Pack Built in This Repo
+
+Executable fixture:
+
+- `packages/conformance-tests/fixtures/industrial/industrial-humanoid-shipyard-safety-pack.v1.json`
+
+Guide:
+
+- `docs/guides/industrial-humanoid-shipyard-safety-pack.md`
+
+Sprint plan:
+
+- `docs/roadmaps/industrial-humanoid-shipyard-safety-sprint.md`
+
+## Proposed Message
+
+SINT does not compete with Persona's robot stack. It complements it as the
+policy, approval, and evidence layer for industrial humanoid deployments. For
+shipyard welding, SINT can prove that hot-work permits, safety-controller state,
+simulation receipts, supervisor approvals, and operator evidence were present
+before the robot acted.
+
+## Suggested Next Conversation
+
+Ask for one narrow integration surface:
+
+- one ROS 2 or controller action for weld start
+- one safety PLC or permit signal
+- one simulator receipt or offline program hash
+- one supervisor approval event
+
+Then map that path into a SINT `request -> decision -> receipt -> execution`
+trace.
+
+## Public References
+
+- Persona AI: `https://persona.ai/`
+- Persona AI RaaS brochure: `https://persona.ai/wp-content/uploads/PAI-RaaS-Brocure-04-2026-v1.pdf`
+- HD Hyundai / Persona AI shipbuilding automation announcement: `https://www.prnewswire.com/news-releases/hd-hyundai-and-persona-ai-sign-agreement-to-deploy-humanoid-welding-robots-for-shipbuilding-automation-302449258.html`
+- ABS / Persona AI shipyard robotics announcement: `https://www.businesswire.com/news/home/20250923457571/en/ABS-and-Persona-AI-Partner-to-Bring-Humanoid-Robotics-to-Shipyards-Advancing-Safety-and-Productivity`
@@ -0,0 +1,121 @@
+# Industrial Humanoid Shipyard Safety Pack
+
+This guide defines a SINT safety and evidence pack for industrial humanoids
+working in shipyards, with emphasis on welding, hot work, confined-space entry,
+inspection, material handling, and remote survey evidence.
+
+The target deployment shape matches companies building humanoid robots for
+heavy industry: a robot may inspect a hull block, enter a shared work zone,
+start a welding arc, grind material, lift parts, or generate evidence for a
+supervisor and surveyor. SINT sits before those actions as the policy gateway.
+
+Executable fixture:
+
+- `packages/conformance-tests/fixtures/industrial/industrial-humanoid-shipyard-safety-pack.v1.json`
+
+Conformance test:
+
+- `packages/conformance-tests/src/industrial-humanoid-shipyard-safety-pack-conformance.test.ts`
+
+## Scope
+
+The pack covers one shipyard block:
+
+- industrial humanoid welder
+- welding supervisor
+- fire watch
+- safety PLC
+- remote surveyor
+- shipyard operator
+- welding torch, grinder, inspection camera, and material gripper
+
+SINT does not certify welding procedures, shipyard safety, robot hardware,
+classification-society acceptance, or OSHA compliance. It provides policy
+decisions, receipts, and evidence records that can support those processes.
+
+## Policy Templates
+
+| Template | Resource | Default tier |
+| --- | --- | --- |
+| Hot-work welding | `humanoid://shipyard/robot/*/tool/welding-torch/*` | T3 |
+| Confined-space entry | `humanoid://shipyard/robot/*/workspace/*/confined-space/enter` | T3 |
+| Material handling | `humanoid://shipyard/robot/*/material/lift` | T2 |
+| Visual inspection | `humanoid://shipyard/robot/*/inspection/*` | T0 |
+
+Hot-work welding requires:
+
+- hot-work permit
+- fire-watch readiness
+- fume extraction online
+- safe gas-atmosphere context
+- simulation receipt bound to the submitted welding program digest
+- fresh safety-controller context
+
+## Required Scenarios
+
+The executable pack covers:
+
+- visual inspection allowed as T0 observe
+- welding denied without a hot-work permit
+- welding denied when fire watch is not ready
+- welding denied when fume extraction is offline
+- confined-space entry denied when gas atmosphere is unsafe
+- bystander in weld zone escalates to T3
+- simulation receipt mismatch denied before arc start
+- valid weld start escalates for supervisor approval
+- material lift over envelope denied
+- e-stop during weld rolls back unconditionally
+
+## ABS-Style Evidence Export
+
+The fixture defines an evidence export profile for remote survey, audit replay,
+data-quality review, and incident reconstruction. Required fields include:
+
+- site and vessel block identifiers
+- robot and work-order identifiers
+- weld procedure specification reference
+- hot-work permit and fire-watch operator identifiers
+- gas monitor calibration reference
+- fume extraction status
+- simulation receipt digest
+- policy decision and tier
+- hash-chain fields: `eventHash`, `previousHash`
+
+Data quality rules require clock sync, calibrated sensors, and program digest
+binding between simulation and execution.
+
+## Persona AI Fit
+
+Persona AI publicly describes industrial humanoids for shipyards, energy,
+construction, and manufacturing, including human coworker safety and
+supervisor/bystander coordination. This pack translates that operating model
+into SINT controls:
+
+- every welding or confined-space action is pre-gated
+- every denial/escalation has a receipt
+- every simulation-to-execution drift case is blocked
+- every incident can be reconstructed through hash-chained evidence
+- every support claim stays inside evidence and runtime-governance boundaries
+
+## Verification
+
+Run the targeted pack:
+
+```bash
+pnpm --filter @pshkv/conformance-tests exec vitest run src/industrial-humanoid-shipyard-safety-pack-conformance.test.ts
+```
+
+Run all fixture contracts:
+
+```bash
+pnpm --filter @pshkv/conformance-tests test:fixtures
+```
+
+## Exit Criteria
+
+- Hot-work, fire-watch, fume extraction, atmosphere, bystander, simulation,
+  supervisor-approval, load-envelope, and e-stop cases are executable.
+- FMEA rows reference executable scenario IDs.
+- Evidence export includes data-quality rules.
+- Certification language remains support-only unless a qualified assessor signs
+  off.
@@ -54,12 +54,14 @@ limits, and tamper-evident audit evidence before execution.
 - Good-first-issues board: [Community/Starter Board](./community/good-first-issues-board.md)
 - Collaboration reply playbook: [Community/Replies](./community/open-source-collaboration-replies.md)
 - Physical AI runtime safety working group: [Community/Working Group](./community/physical-ai-runtime-safety-working-group.md)
+- Industrial humanoid shipyard safety pack: [Guide](./guides/industrial-humanoid-shipyard-safety-pack.md)
 - OWASP Agentic Landscape submission packet: [Community/OWASP Packet](./community/owasp-agentic-landscape-submission.md)
 - EU AI Act mapping: [Compliance/EU AI Act](./compliance/eu-ai-act-mapping.md)
 - ISO 13482 alignment: [Compliance/ISO 13482](./compliance/iso-13482-alignment.md)
 - Formal threat model: [Security/Formal Threat Model](./security/formal-threat-model.md)
 - MITRE ATLAS candidate mappings: [Security/MITRE ATLAS](./security/mitre-atlas-agent-technique-mappings.md)
 - Agent commerce governance profile: [Spec](./specs/agent-commerce-governance-profile-v1.md)
+- Persona AI shipyard safety brief: [Community Brief](./community/persona-ai-shipyard-safety-brief.md)
 - NIST submission bundle report: [Report](./reports/nist-submission-bundle.md)
 - Latest security bulletin: [May 2026](./security-bulletins/2026-05.md)
 

@@ -75,6 +75,13 @@ SINT can govern task-market and machine-payment workflows before work or money
 moves: task creation, bids, claims, benchmark proof submission, worker
 selection, settlement release, and x402-style payment permits.
 
+### 7. Industrial humanoid safety evidence
+
+SINT can govern industrial humanoids before they weld, grind, lift, enter
+confined spaces, or move through shared work zones. The message is not
+"certification replacement"; it is "policy receipts and runtime evidence for
+supervisors, safety teams, surveyors, and post-incident review."
+
 ## Proof points to emphasize
 
 - Apache-2.0 licensed
@@ -98,6 +105,9 @@ selection, settlement release, and x402-style payment permits.
 - "industrial AI safety"
 - "agent commerce governance"
 - "x402 policy enforcement"
+- "industrial humanoid safety"
+- "shipyard robotics governance"
+- "hot-work policy receipts"
 - "tamper-evident evidence ledger"
 - "open protocol and reference stack"
 - "real-world consequences"