tool-failure-diagnostic

To handle tool execution failures automatically, you can use a diagnostic loop that captures stderr and exit codes, audits execution context, and applies corrective actions like retries or configuration updates for autonomous recovery.

A diagnostic loop for autonomous recovery is a process that converts tool failures into diagnostic events by capturing symptoms, generating failure hypotheses, and applying corrective actions like patches or config updates before verifying the fix.

To troubleshoot transient or configuration-related tool failures, audit the working directory, check dependency existence and filesystem permissions, then generate a hypothesis to determine if the issue requires a retry, configuration update, or patch.

Yes, you can automatically verify a corrective action by re-running the original failed command after applying the fix, which confirms whether the transient, configuration, or logical failure has been successfully resolved.

Yes, this autonomous recovery approach works for missing dependencies like the gh command by auditing context for dependency existence, classifying the failure, and automatically attempting to install or fix the missing component.

The diagnostic loop for tool failure recovery requires scripting and diagnostic capabilities to function, and it is limited to failures where context audit, hypothesis generation, and corrective actions like retries or patches can logically resolve the issue.