@ArizePhoenix: This week in Phoenix - feedback gets more visible and the agent gets more capable: Server-side bash for PXI subagents (…
Summary
This week's Phoenix update adds server-side bash for PXI subagents with sandboxed execution and built-in GraphQL access, improving feedback visibility and agent capabilities.
View Cached Full Text
Cached at: 06/25/26, 12:08 AM
This week in Phoenix - feedback gets more visible and the agent gets more capable:
Server-side bash for PXI subagents (beta) - subagents now run a sandboxed bash tool with phoenix-gql built in: same GraphQL access as the main agent, same mutation gate, outbound network off by default behind the SSRF guard.
Similar Articles
@ArizePhoenix: When debugging agents, you need to get to the problematic issues QUICKLY! Phoenix now comes with quick list-details nav…
Arize Phoenix adds quick list-details navigation and resizable drawers for faster agent conversation debugging, plus VIM support in sessions.
@ArizePhoenix: Something we’ve been playing with and liking a lot: Give every coding agent its own observability stack. Because Arize …
Arize Phoenix enables local-first, air-gapped observability for coding agents, allowing each agent to have its own traces, evals, and feedback loop for self-verification.
agent-shell 0.55 updates
agent-shell 0.55 adds support for new AI agents like CodeBuddy, Hermes, and Kimi Code, emphasizing vendor-neutral ACP. The update also discusses ecosystem changes and calls for community support.
@thsottiaux: We are continuing to invest in making agents work better on Windows. Highly recommend reading David's engineering post …
OpenAI is improving agent support on Windows by implementing a custom sandbox for Codex, addressing OS-level isolation challenges to ensure safe and efficient operation.
@ByteMohit: https://x.com/ByteMohit/status/2063493300884246598
A detailed technical post about building AgentForge, an open-source agent harness in Python, covering components like session runtime, tool contracts, approval layers, and persistence, emphasizing that agents are defined by their runtime, not just the model.