Skills synced from claude.ai are now namespaced under anthropic-skills: so they stop silently overriding or being blocked by local skills and commands with the same short name
What
- Skills that come from account sync (claude.ai) can now be referenced and stored under a namespaced key like
anthropic-skills:<name>. A helper strips this prefix when looking the skill up, and a synced skill whose bare (short) name collides with a local or plugin command is no longer simply dropped — it stays usable under its namespaced name. - When Claude Code builds its
/contextinventory of tools and skills, a synced skill's bare/unqualified name is no longer treated as claiming that name slot, so it can't silently take over an unqualified name that a locally defined skill or command also uses. - Name matching for tools and skills now handles synced skills as a special case: instead of using the generic name-suffix matching heuristic, it matches against the skill's own declared aliases.
Why
Before this, a skill synced from your claude.ai account could quietly take over a name also used by a skill or command defined locally, either shadowing it or getting dropped outright. Namespacing synced skills and changing how their names are matched means both the synced version and a same-named local skill or command can coexist and be referenced correctly.
The entry above is what we published on the day. These lines were added later, as Anthropic's own pages caught up, and they sit beside the original rather than replacing it.
5. **Trim context on the server, or compact in a shape that carries no stale thinking:** If your code truncates or summarizes older turns on the client, the simplest fix is to move that work to server-side [compaction](https://platform.cla…models/fable-5-1/migration-guide see the edit
On Claude Fable 5.1, the API accepts a replayed thinking block only while the `system` prompt, `tools`, and messages that preceded it are unchanged. See [Keeping the prefix unchanged](https://platform.claude.com/docs/en/build-with-claude/p…build-with-claude/thinking-troubleshooting see the edit
Anthropic's documentation has since written up /context, on thinking-troubleshooting.