Alloy 0.7.1 (plus 0.7.0): better results, cleaner tooling, and code execution support

[striking]

Alloy has two fresh releases:

• 0.7.1 — introduces Alloy.Result, a typed, Access-compatible result struct returned by Alloy.run/2 and
  Alloy.Agent.Server.chat/3.
• 0.7.0 — adds Anthropic code execution support (code_execution: true), structured tool results, and improved middleware/tool
  lifecycle semantics.

Why this matters:

• Cleaner and safer result handling in app code
• Better support for programmatic tool output flows
• Easier path to richer agent behavior without changing Alloy’s minimal core philosophy

Links:

• Changelog: alloy/CHANGELOG.md at main · alloy-ex/alloy · GitHub
• Hex: alloy | Hex
• Docs: alloy v0.7.1 — Documentation
• GitHub: GitHub - alloy-ex/alloy: Model-agnostic agent harness for Elixir · GitHub

As always, feedback and sharp edges are very welcome.