Code of Conduct

1) Client commitments

• Outcome‑first. We set measurable targets with the client (e.g., cycle time, cost per unit, yield) and review them weekly.
• Plain‑spoken communication. No hype. We explain trade‑offs, risks, and alternatives in clear language.
• Confidentiality. We protect non‑public information, sign NDAs, and minimize data exposure on a need‑to‑know basis.
• Ownership. Clients own their data and custom code produced under contract. We avoid proprietary lock‑in and document handover thoroughly.
• Right to decline. We may decline work that violates this code (e.g., surveillance of workers, deception, or unsafe systems).

2) Security & privacy

• Least privilege & segregation. Access is time‑boxed, role‑scoped, and revoked on project completion.
• Data minimization. We ingest only the data needed for the task and retain it only as long as necessary.
• Encryption. Data in transit and at rest is encrypted with industry‑standard algorithms; secrets are never stored in code repositories.
• Third‑party services. We review data handling terms and opt out of training on client data in public models. Where managed services are used, we isolate tenancy.
• Auditability. We maintain change logs, access logs, and a simple incident log for each engagement.
• Incident response. Suspected breaches are escalated within hours, with containment, remediation, and client notification procedures.

3) AI, robotics & automation safeguards

• Human‑in‑the‑loop for impact. Automated decisions that affect safety, employment, compliance, or customer outcomes require human review until proven safe.
• Model risk management. We document data sources, failure modes, and monitoring plans; we design safe fallbacks for outages or drift.
• Bias & fairness. We test for unwanted bias where decisions affect people. We do not build systems intended to discriminate.
• Safety in robotics. We follow relevant safety standards and implement interlocks, e‑stops, and safe states; changes are validated on hardware.
• Explainability & traceability. We prefer deterministic pipelines, clear prompts, and traceable retrieval to enable audits.
• No dark patterns. We do not design deceptive UX or impersonate humans with automation.

4) People & workplace

• Respect & inclusion. Zero tolerance for harassment or discrimination. We use inclusive language and create accessible interfaces.
• Operator empathy. We involve the people who do the work in design and testing; training is part of the deliverable.
• Transparent change. We communicate what changes, why it changes, and how success will be measured.

5) Integrity & compliance

• Anti‑corruption. No bribes, kickbacks, or undisclosed incentives. Gifts are modest and disclosed.
• Conflicts of interest. We disclose potential conflicts and recuse where appropriate.
• Regulatory awareness. We design with applicable laws and standards in mind (privacy, product safety, export controls) and raise issues early.
• Truthful marketing. We do not name‑drop or exaggerate results; case studies remain anonymized unless clients opt in.

6) Delivery standards

• Small, reversible steps. Thin slices to production with the ability to roll back safely.
• Observability. Dashboards and alerts for data quality, latency, and errors are part of the MVP.
• Documentation & handover. We provide runbooks, diagrams, and checklists. Clients must be able to operate without us.
• Quality management. Defects, incidents, and learnings feed a continuous improvement loop.

7) Sustainability & community

• Pragmatic sustainability. We favor efficient compute, right‑sized hardware, and re‑use where possible.
• Local impact. We support communities where we work and share practical knowledge with operators.

8) Reporting & enforcement

Report concerns or violations to conduct@sektus-technologies.com. We review within a reasonable timeframe and respond with findings and actions.

• Concerns can be raised anonymously; retaliation is prohibited.
• Violations may result in removal from projects or termination of collaboration.

This document is not legal advice. It may evolve; the current version supersedes prior versions. Last updated: August 17, 2025.