Disclaimer

1. No Warranty

SkillsWorkbench is provided “AS IS” and “AS AVAILABLE” without any representation or warranty of any kind, whether express, implied, or statutory. To the fullest extent permitted by applicable law, we expressly disclaim all warranties including but not limited to:

• Merchantability — that the Service is fit for ordinary use
• Fitness for a particular purpose — that the Service meets your specific requirements
• Non-infringement — that the Service does not infringe any third-party rights
• Accuracy or completeness of any information produced by or through the Service
• That the Service will be uninterrupted, error-free, secure, or free from viruses or harmful components
• That defects in the Service will be corrected

2. AI Output Disclaimer

The Skill drafts, stress-test scenarios, evaluation questions, and other content generated through SkillsWorkbench are produced by large language models (LLMs), specifically Anthropic's Claude API. You acknowledge and agree that:

• AI outputs may be incorrect. Language models can generate plausible-sounding but factually wrong, outdated, or misleading content. We make no representation that any generated output is accurate, complete, or appropriate for your use case.
• AI outputs may be inconsistent. The same prompt may produce materially different outputs across sessions. Skill behaviour in production may differ from behaviour observed during testing on this platform.
• AI outputs may reflect biases. Language models may exhibit biases embedded in their training data. Skills generated on this platform may inadvertently reflect or amplify those biases.
• AI outputs may be harmful. Generated Skills could, if deployed carelessly, produce responses that are inappropriate, harmful, or offensive to end users. You bear sole responsibility for reviewing content before deployment.
• We do not review your outputs. We have no visibility into, and take no responsibility for, the Skill prompts you design or the way you deploy them.

You are solely responsible for all Skill content you create using this Service and for any consequences of deploying that content in a production AI system.

3. Not Professional Advice

Nothing produced by SkillsWorkbench — including Skill drafts, evaluation results, stress-test outputs, or any in-app guidance — constitutes legal, medical, financial, psychological, compliance, or any other form of professional advice. Specifically:

• Legal: No output from this Service should be construed as legal advice. Consult a qualified legal professional before deploying AI systems in regulated legal contexts.
• Medical / Healthcare: No output constitutes medical advice, diagnosis, or treatment recommendations. AI-assisted healthcare applications must be validated by qualified medical professionals and comply with applicable clinical and regulatory standards.
• Financial: No output constitutes investment, tax, or financial advice. Financial AI applications are subject to regulatory requirements that vary by jurisdiction and must be assessed by qualified professionals.
• Compliance: References to regulatory frameworks (SOX, HIPAA, GDPR, ISO, etc.) in any generated content are illustrative only. They do not constitute a compliance assessment or legal opinion. Seek qualified professional advice for regulatory compliance.

4. Third-Party Dependencies

SkillsWorkbench depends on third-party services including the Anthropic Claude API and Google Firebase. We:

• Cannot guarantee the continuous availability or performance of those services
• Are not responsible for changes to third-party APIs, pricing, or policies that affect the Service
• Do not control the data processing practices of third-party providers — their own terms and privacy policies govern your data as processed by them
• Are not affiliated with, endorsed by, or acting as an agent of Anthropic PBC, Google LLC, or any other third party

Anthropic may change its model behaviour, safety filters, or API capabilities at any time, which may affect how Skills you design behave in production. We are not liable for such changes.

5. Service Availability

We do not guarantee that the Service will be available at any particular time. The Service may be:

• Temporarily unavailable due to scheduled maintenance, infrastructure failures, or events beyond our control
• Modified, suspended, or discontinued at any time with or without notice
• Subject to rate limits, usage quotas, or throttling imposed by us or our infrastructure providers

We are not liable for any losses or damages resulting from Service downtime or unavailability.

6. Limitation of Liability

To the maximum extent permitted by applicable law, SkillsWorkbench and its officers, directors, employees, contractors, and agents shall not be liable for:

• Any direct, indirect, incidental, special, exemplary, punitive, or consequential damages
• Loss of profits, revenue, data, goodwill, or business opportunities
• Damages arising from reliance on AI-generated content
• Damages arising from the deployment of Skills in production systems
• Regulatory fines, penalties, or enforcement actions resulting from the use or misuse of the Service
• Harm caused to third parties by AI systems built using Skills created on this platform

Our aggregate liability to you shall not exceed the greater of (a) USD $100 or (b) amounts paid by you to us in the 12 months preceding the event giving rise to the claim, even if we have been advised of the possibility of such damages.

Some jurisdictions do not permit the exclusion or limitation of certain types of damages, so some of the above may not apply to you.

7. Regulatory Compliance

If you intend to use Skills created on this platform in regulated industries or applications, you are solely responsible for ensuring compliance with all applicable laws, regulations, and standards, including but not limited to:

This list is illustrative, not exhaustive. Regulatory landscapes evolve rapidly. We do not monitor applicable regulations on your behalf and are not responsible for your compliance obligations.

8. Production Use Warning

Before deploying any Skill prompt created using SkillsWorkbench in a production AI system, you should at a minimum:

1. Conduct independent testing with real-world inputs representative of your user population
2. Perform a human review of generated outputs for accuracy, tone, and safety
3. Implement appropriate human oversight and escalation paths for high-stakes decisions
4. Conduct a risk assessment appropriate to the sensitivity of the application
5. Obtain legal and regulatory review if deploying in a regulated industry
6. Establish monitoring and feedback loops to detect performance degradation over time
7. Comply with Anthropic's usage policies for the Claude API

SkillsWorkbench provides a design and testing environment. It does not replace rigorous engineering, safety review, or compliance processes required for production AI systems.

9. Contact

If you have questions about this Disclaimer or believe you have identified a safety issue with the Service, please contact:

SkillsWorkbench
Email: contact@skillsworkbench.com