Security and privacy boundaries

What Fygment does and does not expose in public launch docs, support reports, admin views, imports, MCP, and billing metadata.

Audience: Workspace operators, support reviewers, and users checking trust boundaries.
Updated: 2026-05-20

The public launch trust model is metadata-minimizing: keep customer Skill content, package bodies, prompts, secrets, tokens, and raw local paths out of admin, support, analytics, and billing surfaces.

Data minimization rules

Platform-admin views are for support and aggregate operations. They must not expose customer Skill bodies, package contents, proposal patches, feedback bodies, secrets, tokens, prompts, stdout, stderr, or raw local paths.

Billing metadata should stay workspace-scoped and minimal: workspace id or slug, initiating user id, plan key, status, quantity, and mode-style fields only.

No enterprise overclaim

Compliance and marketplace claims stay gated

Public docs must not present SSO, SCIM, SOC 2, ISO 27001, on-prem, SIEM export, broad security scanning, or public marketplace workflows as live launch features.

Related docs

Troubleshooting
Start with safe, non-sensitive checks. Fygment support reports should capture bounded diagnostics, not package bodies, secrets, raw prompts, or local paths.
Billing and launch mode
Public launch uses a workspace effective tier: Solo remains free for the local creator loop, Advanced / Teams unlocks paid collaboration and external import actions, and Enterprise stays support-managed.
Review and approve Skills
Governance is the difference between a shared Skill library and a loose pile of files. Review decisions decide what can be installed and trusted.