The value tools split by scope — one flag in one Environment, or every flag in a Config in a single call — and each of those writes takes a reason alongside the new value. Both the original and the v2 setting-value tools ship, and the v2 pair is the one that replaces Targeting Rules, so a client written against the older shape keeps working. One tool does no flag work at all: update-sdk-documentation downloads ConfigCat SDK documentation from a URL you pass it, for the point where the next step is code rather than configuration.
A cover over the ConfigCat management endpoints — 79 tools spanning Organizations, Products, Configs, Environments, feature flags, Segments, Tags, Webhooks, Permission Groups and audit logs.
- Flag values read and written per Environment, in the original shape and in the v2 shape that carries Targeting Rules
- A reason string on every value write, including the batch call that updates a whole Config at once
- Zombie (stale) flags for a Product, narrowed by age in days, by environment or by tag
- Code references for a flag, and the SDK Key for a Config in a given Environment
- Audit log items for a Product or for an Organization, over a date range you set
- Products, Configs, Environments, Segments, Tags and Webhooks created, updated and deleted
- Members invited, moved between Permission Groups, and removed from a Product or an Organization
A ConfigCat account with API credentials in CONFIGCAT_API_USER and CONFIGCAT_API_PASS. Nothing has to be looked up first: list-organizations and list-products return what the account can see, and the rest of the catalogue works from the ids those two hand back.
One command plus a key — npx -y @configcat/mcp-server, then supply credentials