execute-task

Execute browser tasks from natural-language descriptions and capture evidence traces.

4|1|Updated Feb 3, 2026

Search Tags:#automated-testing #execute #browser-automation #evidence-capture #task-trace #webbench #browser-navigation

Ready to run

npx skills add https://github.com/lespaceman/athena-workflow-marketplace --skill execute-task-lespaceman

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: execute-task
Source: https://github.com/lespaceman/athena-workflow-marketplace/tree/main/plugins/web-bench/skills/execute-task
Command: npx skills add https://github.com/lespaceman/athena-workflow-marketplace --skill execute-task-lespaceman

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Automates browser-based task execution and evidence capture for WebBench benchmarks.

Core Features & Use Cases

Interprets natural-language task descriptions to determine required browser actions.
Plans and executes deterministic browser steps, capturing traces and final evidence (screenshots, snapshots, and final URL).
Generates a structured execution trace including timing, results, and any blockers or errors.
Use Case: You want to benchmark a site's performance by navigating to a page, executing a script, and recording outcomes.

Quick Start

Provide a single natural-language task description to start a WebBench execution.

Frequently Asked Questions about execute-task

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I automate browser-based task execution for WebBench benchmarks?▼

Browser automation for WebBench benchmarks is executed by interpreting natural-language task descriptions, planning deterministic browser steps, and capturing an execution trace with results and blockers. It automates the entire process from navigation to evidence capture.

What evidence does browser automation capture during web task execution?▼

Evidence capture during web task execution records screenshots, page snapshots, and the final URL. It generates a structured execution trace that includes timing metrics, step results, and any encountered blockers or errors.

Do I need a specific environment to run automated browser navigation and evidence capture?▼

Automated browser navigation requires a calling context with browser automation capabilities. You provide a natural-language task description, and the environment must support executing deterministic browser actions and capturing traces.

Can browser automation scripts handle login pages or destructive actions?▼

Browser automation enforces guardrails to avoid login, credential guessing, or destructive actions. It is designed to handle common obstacles during navigation but explicitly blocks tasks requiring authentication or data modification.

Why is my automated web task execution not completing successfully?▼

Automated web task execution may fail due to common navigation obstacles or enforced guardrails. The generated execution trace logs specific blockers, errors, and timing details to help diagnose why the browser actions could not complete.