Run software engineering tasks with OpenHands
Use OpenHands to run supervised software-engineering agents that inspect repositories, edit code, execute commands, and return reviewable implementation evidence.
npx skills add agentskillexchange/skills --skill run-software-engineering-tasks-with-openhands
Use OpenHands when an operator needs a repeatable software-engineering agent workflow for repository tasks that require code inspection, edits, command execution, and reviewable handoff. The workflow is to start an OpenHands local GUI, CLI, SDK, or hosted runtime, connect a repository and model provider, define the implementation task, supervise tool use, and review the produced diff, logs, and reasoning before merging.
What this skill actually does
Invoke this instead of asking a general chatbot for code when the work needs an agent loop with workspace access, terminal execution, browser or API feedback, and an auditable path from task brief to patch. The scope boundary is supervised software-engineering task execution through the OpenHands runtime; it is not a generic model provider, IDE replacement, or broad AI-development platform card.
Inputs are a target repository or workspace, model/provider credentials, task instructions, runtime permissions, and optional evaluation or benchmark harnesses. Outputs are code changes, command logs, failure traces, task summaries, and implementation evidence suitable for human review.