Reference
FAQ
Common questions about Cendis — pricing, security, supported tools, asset formats, and how it compares to building your own registry.
What is Cendis?
Cendis is a governed registry for AI coding assets — skills, commands, templates, and configuration files for AI coding agents. You publish once, your projects pull, and you have a complete audit trail of which version runs where. Claude Code is the first supported agent; Cursor and Codex support are on the roadmap.
Which AI tools does Cendis support?
Claude Code is fully supported today — the CLI distributes skills, commands, and CLAUDE.md files into .claude/skills/ and .claude/commands/.
Cursor and Codex support are on the roadmap. The registry is designed to be tool-agnostic: when new agents ship, adding a target is as simple as updating the asset’s target_path — no workflow changes required.
How is this different from putting assets in a Git repo?
A Git repo gives you version control. Cendis adds:
- SHA-verified pulls — files are checksummed before being written; tampering breaks the pull
- Approval queue — new versions can require sign-off before becoming
published - Drift detection —
cendis statusshows what’s outdated, missing, or hand-edited - Audit log — who published, who approved, who pulled, and when
- Multi-tool distribution — one asset, multiple targets (skill ships to both Claude Code and Cursor)
Most teams that try the Git-repo approach end up rebuilding 60% of Cendis poorly. We’ve already built it.
What’s the pricing?
- Trial — free, time-limited, full feature access
- Team — $99/workspace/month
- Growth — $249/workspace/month
- Enterprise — custom
You pay per workspace, not per seat. A workspace is a governance scope (typically one product line or one team). Most orgs start with one and add more as they grow.
See pricing for the live numbers and feature matrix.
Is Cendis open source?
The CLI is open source. The registry, dashboard, and approval workflow are commercial. Self-hosting the full stack is on the Enterprise tier.
How does Cendis handle secrets?
Cendis is a registry for governance assets — not a secret manager. Tokens, API keys, and credentials should never be committed to a Cendis asset.
The CLI itself uses an environment variable (CENDIS_TOKEN) for authentication. The token is read from your shell or CI environment, never written to disk by the CLI, and never logged. See Authentication for details.
Can Cendis run in our private network?
Yes — Enterprise customers can self-host the registry behind their VPN. The CLI talks to the registry over HTTPS; point it at your internal hostname with the --registry flag (or in cendis.yaml).
What happens if I lose my token?
You can’t recover a Personal Access Token — they’re shown once at creation. Revoke the lost token from the dashboard and create a new one. See Rotate tokens.
How do I publish my first asset?
Asset publishing currently runs through the dashboard. CLI publish (cendis publish) is on the roadmap.
In the dashboard:
- Go to Assets → New Asset
- Choose the type (skill, command, template, rules file)
- Upload the file or paste the content
- Set the slug, version, and changelog
- Choose targets (Claude Code, Cursor, etc.)
- Submit — enters the approval queue if your workspace requires it
Can I version assets independently of each other?
Yes. Each asset has its own semantic version (major.minor.patch). They are not coupled. Publishing v2.0.0 of skill/code-review has no effect on command/deploy-check.
Does Cendis work offline?
Pulled assets are written to disk and work offline indefinitely — your AI agents read them locally. The CLI itself needs network for pull, list, status, and import because those talk to the registry.
For air-gapped environments, the Enterprise tier supports a local registry mirror.
How do I migrate from an existing in-repo collection?
Use cendis import:
cendis import --source ./ai-assets --phase <phase-uuid>
The importer scans the directory, infers asset types from folder names (skills/, commands/, templates/, rules/), and bulk-uploads to your registry as draft versions. Review and publish from the dashboard.
What file formats are supported?
For asset content: .md, .yaml, .yml, .json, .txt. Most assets are markdown.
The CLI itself reads cendis.yaml (the project config) — YAML, not JSON, because tokens are interpolated from environment variables (token: ${CENDIS_TOKEN}).
How do I get help?
- Docs: you’re reading them
- Email: justin@cendis.ai
- Discord: linked from the community page (coming soon)
How do I report a bug or request a feature?
Open an issue on the Cendis GitHub — we triage every one.
Next steps
- CLI Quickstart — get the CLI running in five minutes
- Concepts — understand the governance model
- Authentication — manage tokens for local + CI
Last updated: 2026-04-17