OpenClaw Signals: Modularizing Execution, Native Telemetry, and Async I/O

Refactor

Executor Modularization: Process, HTTP, and SSH Separation

PR #329 splits the monolithic 800-line system executor into dedicated modules for process, HTTP, and SSH. This isn't just hygiene; it's a necessary step to get concurrency under control and stop the bleeding on edge cases like port exhaustion.

Fix

Standardized Env Parsing for Containerized Executors

If you're running OpenClaw in a containerized environment, this matters. PR #328 standardizes how the system parses environment variables for executors, cleaning up a mess where keys were lowercase in some configs but case-sensitive in others.

Tooling

Better Diagnostics: Execution History Drawer

Issue #1560 introduces a history inspection drawer specifically for failed executions. No more digging through raw logs to figure out why a specific agent run died; it’s a direct UI shortcut to the root cause.

Observability

Native System Monitoring Replaces External Sidecars

Up until now, tracking system metrics in OpenClaw meant hacking together your own sidecar. This PR moves system monitoring directly into the core platform, giving you built-in visibility into resource consumption out of the box.

Enhancement

Async Input Processing Unblocks Large File Transfers

Writing agents that need to handle large uploads just got less painful. PR #326 decouples the input handling from the main event loop, which should eliminate the UI freezes we saw when agents processed files over 10MB.

Roadmap

Feature Proposal: Salting to Prevent Infinite Loops

This is a feature request for 'salted generation'—essentially seeding the agent's output with entropy to prevent drift in long-running loops. It’s a niche fix, but vital for keeping deterministic systems from freezing up.