Trust & Safety
How MyClaude enforces content policies, automated scanning, MCS certification, and dispute resolution to maintain marketplace integrity.
MyClaude enforces content policies, automated scanning, and human review to maintain marketplace trust.
Every product on MyClaude passes through controls at publish time and is subject to ongoing review. This page explains how those controls work.
Content Scanning
When a creator runs myclaude publish, the CONDUIT pipeline runs myclaude scan against the full product bundle before any listing is created. Scanning is not optional and cannot be bypassed.
| Scan type | What is checked |
|---|---|
| Secrets detection | API keys, tokens, private keys, credentials — flagged by pattern matching and entropy analysis |
| Malicious code patterns | Shell injection, network exfiltration, unauthorized file access, data exfiltration payloads |
| Prohibited content | Content that violates the Content Policy |
| Manifest integrity | vault.yaml schema validity, required fields, version format, category classification |
A product that fails any scan is rejected before listing. The creator receives a specific, actionable error message. Scan results are not published publicly.
Passing automated scanning does not guarantee a product is free of all issues — it establishes a baseline. Additional controls apply after listing.
MCS Certification
MCS (MyClaude Certified Standard) is a quality badge awarded to products that pass both automated review and a manual human review.
| Level | Controls |
|---|---|
| Listed (no badge) | Automated scan passed |
| MCS Certified | Automated scan + manual human review |
MCS-certified products:
- Appear higher in search results
- Display a verification badge on the product page and in CLI output
- Have passed a human review of code quality, documentation completeness, and content policy compliance
Creators can request MCS review after publishing by contacting support. Review time varies by product complexity; expect 3-7 business days at current scale.
Prohibited Content
The following categories of content are prohibited on MyClaude and will be rejected at scan time or removed after listing if identified later. Full definitions are in the Content Policy.
| Category | Examples |
|---|---|
| Malicious code | Any code designed to harm, surveil, or exploit systems or users |
| Credential theft | Products that capture, transmit, or store user credentials without disclosure |
| Data exfiltration | Products that send user data to third parties without disclosure |
| CSAM | Zero tolerance; immediately reported to authorities |
| Deceptive products | Products that misrepresent their function, category, or capabilities |
| Plagiarism | Products that copy another creator's work without license or attribution |
| Illegal content | Content that violates applicable law |
Post-Listing Monitoring
Automated scanning at publish time is the first line of defense. Additional controls apply after a product is live:
- Community reports — Buyers and creators can report products via the product page. Reports are reviewed within 3 business days.
- Periodic re-scan — Published products are subject to re-scanning if the policy or scan rules are updated.
- Creator account review — Repeated violations result in account suspension.
Dispute Resolution
Disputes between buyers and creators are handled as follows:
| Dispute type | Resolution path |
|---|---|
| Product does not work as described | Contact creator first; escalate to MyClaude support if unresolved within 7 days |
| Unauthorized charge | Contact MyClaude support; Stripe dispute process applies |
| Creator misconduct | Report via product page; MyClaude reviews within 3 business days |
| Creator account impersonation | Report to security@myclaude.sh |
MyClaude acts as a platform intermediary. We do not arbitrate commercial disputes between buyers and creators but will remove products and suspend accounts when policy violations are confirmed.
Refund decisions on paid products rest with the creator, except where law requires otherwise. MyClaude reserves the right to issue refunds and recover platform fees in cases of confirmed fraud or policy violation.
Enforcement Actions
| Violation severity | Action |
|---|---|
| Minor / first offense | Warning + required fix within 7 days |
| Moderate / repeat offense | Product unpublished pending review |
| Severe (malicious code, CSAM) | Immediate account suspension, product removal, authorities notified where applicable |
Creators may appeal enforcement actions by emailing support@myclaude.sh within 14 days of the action. Appeals are reviewed by a team member not involved in the original decision.
Related pages
- Security Model — authentication, authorization, and infrastructure security
- Vulnerability Disclosure — how to report security issues
- Content Policy — full prohibited content definitions