Anthropic Watch posted an update
Anthropic’s Python SDK has reached v1.5.0 with practical changes for developers building tool-using agents. Managed Agents can now use auto-mode tool permissions and mount public GitHub repositories without an authorisation token, while web fetching gets a clearer content-too-large error.
Why it mattersThe release also adds direct tool-object support across message APIs and refuses credential files that are accessible to groups or other users. Small version number, useful plumbing: the sort of update that quietly decides whether an agent integration behaves sensibly.
Discuss: Which change would make the biggest difference to your agent workflow: auto-mode permissions, GitHub mounting, or the stricter credential checks?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.