Coordinate Clawperator eval runs across local-dev and published runtimes.
npx skills add https://github.com/clawperator/clawperator --skill evals-run
Please help me install this Agent Skill. Skill: evals-run Source: https://github.com/clawperator/clawperator/tree/main/.agents/skills/evals-run Command: npx skills add https://github.com/clawperator/clawperator --skill evals-run
Coordinating and running Clawperator evals across local-dev and published runtimes to diagnose preflight failures, compare targets, and triage artifacts.
Run an android-version eval with a chosen agent, model, and runtime to validate the end-to-end workflow.
High-intent search queries and answers about installing and using this skill.
You can replay and rescore eval runs by executing helper scripts that drive emulator and device workflows, generating artifact outputs and run logs to debug preflight failures and validate agent performance across runtime targets.
Triage preflight failures by using the built-in runbook and helper scripts to orchestrate eval runs across local-dev and published runtimes, capturing artifact outputs and run logs to compare targets and validate agent behavior.
Yes, you can run evals across multiple targets using the run_pack_a_android_version_matrix.sh script to drive emulator and device workflows, comparing runtime targets across connected devices for comprehensive end-to-end validation.
Yes, you need to configure the operator setup using the references/evals-run.md runbook, which provides instructions to install and configure the operator and APKs on connected devices before executing the android-version eval workflows.
Comparing local-dev and published runtimes isolates environment-specific issues by running the same android-version evals across both targets, exposing discrepancies in agent behavior and preflight failures through captured artifact logs.
Replay workflows may fail to capture artifacts if the operator and APKs are not properly installed on connected devices, requiring a review of the runbook configuration and the run_android_version_eval.sh script execution environment.
software-mansionBoot an Android AVD and configure adb reverse for Metro connectivity.
open-thoughtsSynthesize evaluation traces, training logs, and GPT-5 assessments into a structured behavioral report.
TrainLCDBoots Android Virtual Devices and configures ADB serials and network for testing workflows.
aaronbassettAutomate Flutter deployment guidance across Android, iOS, Web, and Desktop.
